io.github.JcJamet/ia-qa-toolbox

Est-ce que io.github.JcJamet/ia-qa-toolbox est sûr ?

Au dernier audit MCP Hub, le serveur MCP « io.github.JcJamet/ia-qa-toolbox » est classé « Non encore testé ». Référencé dans l'annuaire. Seul le pentest comportemental fait foi, et il n'a pas encore été exécuté sur ce serveur. Le verdict est réévalué en continu ; consultez la fiche pour l'état à jour.

Méthode : MCP Hub analyse la description de chaque outil (couche 1) et, pour le verdict fort, exécute le serveur dans une micro-VM isolée avec leurres et réseau bloqué (couche 2). On publie des noms et des compteurs, jamais le contenu brut renvoyé par un outil.

Les 152 outils

• ab_test_report≈ 350 tokens

  Generate an A/B test report comparing two prompts or model configurations. Accepts arrays of scores and returns statistical comparison: mean, median, std deviation, winner, and improvement percentage.

  Paramètres, retour & exemple (2)

  Paramètres

  • variant_a · object · requis
    First variant configuration with name and score array
  • variant_b · object · requis
    Second variant configuration with name and score array

  Exemple d'appel

  {"name": "ab_test_report", "arguments": {"variant_a": {}, "variant_b": {}}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"max": "texte", "min": "texte", "mean": "texte", "count": 1, "median": "texte", "winner": "texte", "std_dev": "texte", "variant_a": {}, "variant_b": {}, "recommendation": 1, "improvement_percent": "texte"}}

• analyze_diff_bugs≈ 388 tokens

  Pattern-based diff linter: flags a fixed set of risky shapes in changed code — query-string interpolation (SQL/Cypher/Mongo injection shape), shell interpolation, eval/new Function, empty catch blocks, regex built from a variable, fewer catch blocks than before, and named authorization guards that disappeared. Every finding cites the line that produced it. It does NOT do data-flow analysis: it cannot follow a value to a sink, across functions or files, and an empty result is not a safety verdict (the response lists what it did not analyse). Advisory triage — use a static analyser for a real security gate.

  Paramètres, retour & exemple (3)

  Paramètres

  • context · string · optionnel
    Optional PR title or feature context for better analysis
  • version1 · string · optionnel
    Original code (before changes). If omitted, only the new version is analysed.
  • version2 · string · requis
    New/modified code (after changes)

  Exemple d'appel

  {"name": "analyze_diff_bugs", "arguments": {"version2": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"bugs": [], "disclaimer": "texte", "notAnalysed": [], "overallRisk": "texte", "rulesApplied": 1, "scannedLines": 1, "totalSuggestions": 1}}

• analyze_responses≈ 409 tokens

  Semantically analyze N already-produced model outputs for the SAME task (the MCP counterpart to the LLM Sandbox). Without a reference: computes consensus — pairwise cosine agreement, the most-representative output, and the outlier. With a `reference` (ground truth): also ranks every output by closeness (token cosine + ROUGE-L composite) and names the closest. No LLM, no key, runs locally. The reference mode is gate-able in CI; the consensus/outlier mode is a signal (no reference to fail against). You bring the outputs (2+). For a 2-way head-to-head with structural JSON diff use compare_responses instead.

  Paramètres, retour & exemple (2)

  Paramètres

  • reference · string · optionnel
    Optional ground-truth answer. If set, each output is also ranked by closeness to it and the closest one is named.
  • responses · array · requis
    The outputs to analyze (same task, N models/prompts/versions). Each item is a plain string or { "label": "GPT-4o", "text": "..." }. At least 2 required.

  Exemple d'appel

  {"name": "analyze_responses", "arguments": {"responses": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"count": "texte", "summary": "texte", "consensus": "texte", "reference_ranking": "texte"}}

• base64_decode≈ 138 tokens

  Decode a Base64 string back to UTF-8 text. Use for inspecting Base64-encoded API responses, JWT payload claims, config file values, or attachment data.

  Paramètres, retour & exemple (1)

  Paramètres

  • input · string · requis
    Base64 string to decode

  Exemple d'appel

  {"name": "base64_decode", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"decoded": "texte"}}

• base64_encode≈ 143 tokens

  Encode a UTF-8 string to Base64. Use when you need to embed binary data, multi-line text, or special characters safely inside JSON fields, HTTP headers, or data URIs.

  Paramètres, retour & exemple (1)

  Paramètres

  • input · string · requis
    Text to encode

  Exemple d'appel

  {"name": "base64_encode", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"encoded": "texte"}}

• bias_detect≈ 406 tokens

  Analyse a set of LLM responses generated from the same prompt template but with different demographic variants (gender, origin, age, tone). Returns a bias score (0-100), sentiment analysis per variant, pairwise Jaccard similarity, and a human-readable verdict. No API key needed — runs entirely locally.

  Paramètres, retour & exemple (1)

  Paramètres

  • responses · array · requis
    Array of variant responses to compare for bias

  Exemple d'appel

  {"name": "bias_detect", "arguments": {"responses": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"verdict": "texte", "lengthCV": 1, "biasScore": 1, "sentiments": [], "avgSimilarity": 1, "minSimilarity": 1, "sentimentVariance": 1, "pairwiseSimilarities": []}}

• bm25_score≈ 304 tokens

  Compute BM25 relevance score between a query and one or more documents. BM25 is the industry-standard keyword-based ranking algorithm used in Elasticsearch, OpenSearch, and Weaviate hybrid search. Returns ranked results with normalized scores.

  Paramètres, retour & exemple (5)

  Paramètres

  • b · number · optionnel
    Length normalization factor (default: 0.75)
  • k1 · number · optionnel
    Term frequency saturation (default: 1.5)
  • query · string · requis
    The search query
  • top_k · number · optionnel
    Return top K results (default: all)
  • documents · array · requis
    Array of documents to rank

  Exemple d'appel

  {"name": "bm25_score", "arguments": {"query": "texte", "documents": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"b": "texte", "k1": "texte", "index": "texte", "query": "texte", "results": "texte", "bm25_score": 1, "doc_length": 1, "doc_preview": "texte", "avg_doc_length": 1, "documents_count": 1}}

• build_rag_prompt≈ 389 tokens

  Assemble a complete RAG (Retrieval-Augmented Generation) prompt from retrieved context chunks and a user query. Handles token budgeting, citation numbering, system instruction injection, and source attribution.

  Paramètres, retour & exemple (6)

  Paramètres

  • query · string · requis
    The user question to answer
  • chunks · array · requis
    Retrieved context chunks with .text (required), .source (optional), .score (optional)
  • language · string · optionnel
    Response language instruction (e.g. "French", "Spanish")
  • cite_sources · boolean · optionnel
    Add [1], [2] citation numbers (default: true)
  • max_context_tokens · number · optionnel
    Max tokens for context section (default: 2000)
  • system_instruction · string · optionnel
    Custom system instruction (default: standard RAG grounding instruction)

  Exemple d'appel

  {"name": "build_rag_prompt", "arguments": {"query": "texte", "chunks": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"prompt": "texte", "system_prompt": "texte", "chunks_included": 1, "included_chunks": "texte", "chunks_truncated": 1, "total_tokens_estimate": 1, "context_tokens_estimate": 1}}

• calculate_readability≈ 195 tokens

  Calculate readability scores: Flesch Reading Ease, Flesch-Kincaid Grade Level, Coleman-Liau Index, and Automated Readability Index. Useful for evaluating LLM output quality.

  Paramètres, retour & exemple (1)

  Paramètres

  • input · string · requis
    Text to analyze for readability

  Exemple d'appel

  {"name": "calculate_readability", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"level": "texte", "stats": {}, "coleman_liau_index": 1, "flesch_reading_ease": "texte", "flesch_kincaid_grade": 1, "automated_readability_index": 1}}

• case_convert≈ 199 tokens

  Convert a string between naming conventions: camelCase, PascalCase, snake_case, kebab-case, UPPER_SNAKE_CASE, dot.case, Title Case. Essential for code generation and refactoring.

  Paramètres, retour & exemple (2)

  Paramètres

  • to · string · requis
    Target case: "camel", "pascal", "snake", "kebab", "upper_snake", "dot", "title"
  • input · string · requis
    String to convert (e.g., "myVariableName", "my-css-class")

  Exemple d'appel

  {"name": "case_convert", "arguments": {"to": "texte", "input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"result": "texte", "from_words": "texte", "target_case": "texte"}}

• check_contrast_ratio≈ 224 tokens

  Calculate WCAG 2.1 contrast ratio between two colors. Returns ratio and compliance for AA/AAA normal and large text.

  Paramètres, retour & exemple (2)

  Paramètres

  • background · string · requis
    Background color in hex (e.g., "#ffffff")
  • foreground · string · requis
    Foreground color in hex (e.g., "#333333")

  Exemple d'appel

  {"name": "check_contrast_ratio", "arguments": {"background": "texte", "foreground": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"ratio": "texte", "AA_large": true, "AAA_large": true, "AA_normal": true, "AAA_normal": true, "background": {}, "foreground": {}, "ratio_text": "texte"}}

• color_convert≈ 196 tokens

  Convert a color between HEX, RGB, and HSL formats. Use when translating design tokens between CSS notations, verifying color accessibility, or normalizing color values from user input. Accepts #rrggbb, #rgb, rgb(r,g,b), or hsl(h,s%,l%).

  Paramètres, retour & exemple (1)

  Paramètres

  • input · string · requis
    Color value to convert, e.g. "#ff6b6b", "rgb(255,107,107)", "hsl(0,100%,71%)"

  Exemple d'appel

  {"name": "color_convert", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"b": "texte", "g": "texte", "r": "texte", "hex": "texte", "hsl": "texte", "rgb": "texte", "input": "texte"}}

• compare_models≈ 241 tokens

  Compare 2-5 AI models side by side: context window, pricing, multimodal, reasoning capabilities, and provider. Returns a comparison table with a recommendation based on your use case.

  Paramètres, retour & exemple (2)

  Paramètres

  • models · array · requis
    Array of 2-5 model names (e.g. ["gpt-4o","claude-3.5-sonnet","gemini-2.0-flash"])
  • use_case · string · optionnel · cost | context | reasoning | multimodal | speed
    Optimize recommendation for this criterion

  Exemple d'appel

  {"name": "compare_models", "arguments": {"models": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"rows": "texte", "model": "texte", "use_case": "texte", "recommendation": "texte", "models_compared": 1, "cost_per_1k_total": "texte"}}

• compare_responses≈ 488 tokens

  Compare two ALREADY-PRODUCED outputs (e.g. model A vs model B on the same task) side by side. Returns deterministic metrics (token cosine, ROUGE-L, Jaccard, length/structure deltas, JSON diff) and a verdict. If a `reference` (ground truth) is given, scores each output against it and picks the closer one. If `model` + `api_key` are given, an LLM judge also picks a qualitative winner for the task. No re-execution — you bring the outputs.

  Paramètres, retour & exemple (9)

  Paramètres

  • task · string · optionnel
    The task/prompt both outputs were answering — used by the LLM judge for context
  • model · string · optionnel
    Optional judge model id (BYOK). When set with api_key, an LLM judge picks a qualitative winner.
  • api_key · string · optionnel
    Optional API key for the judge model (BYOK). Used only for the judge call; never stored.
  • label_a · string · optionnel
    Label for output A (e.g. "GPT-4o", "v1.0")
  • label_b · string · optionnel
    Label for output B (e.g. "GPT-5-nano", "v1.1")
  • reference · string · optionnel
    Optional ground-truth / expected answer. If set, each output is scored against it and the closer one wins (deterministic).
  • check_json · boolean · optionnel
    Try to parse as JSON and compare structurally (keys, types, values)
  • response_a · string · requis
    First output (e.g. model A's answer)
  • response_b · string · requis
    Second output (e.g. model B's answer)

  Exemple d'appel

  {"name": "compare_responses", "arguments": {"response_a": "texte", "response_b": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"judge": "texte", "labelA": "texte", "labelB": "texte", "metrics": "texte", "summary": "texte", "verdict": "texte"}}

• consistency_check≈ 308 tokens

  Compare multiple LLM responses to the same prompt and detect inconsistencies using Jaccard word-overlap similarity and fact drift (number comparison). Fast, deterministic, no API key needed. Limitations: relies on surface-level word matching — "Paris is the capital of France" vs "Paris is the French capital" may score low despite semantic equivalence. For true semantic consistency, use run_semantic_tests with embedding mode. Essential for determinism testing.

  Paramètres, retour & exemple (2)

  Paramètres

  • responses · array · requis
    Array of 2+ LLM responses to compare (same prompt, different runs)
  • check_facts · boolean · optionnel
    Check for contradictory numbers/facts across responses (default: true)

  Exemple d'appel

  {"name": "consistency_check", "arguments": {"responses": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"verdict": "texte", "fact_drift": "texte", "avg_similarity": "texte", "response_count": 1, "pairwise_scores": "texte", "fact_contradiction": "texte", "length_variance_percent": "texte"}}

• context_window_check≈ 308 tokens

  Given an array of message objects [{role, content}], estimate total token usage and check if it fits in the target model's context window. Warns about truncation risk.

  Paramètres, retour & exemple (3)

  Paramètres

  • model · string · requis
    Target model name (e.g. gpt-4o, claude-3.5-sonnet)
  • messages · array · requis
    Array of messages (system/user/assistant)
  • max_output_tokens · number · optionnel
    Reserved tokens for output (default: 4096)

  Exemple d'appel

  {"name": "context_window_check", "arguments": {"model": "texte", "messages": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"fits": "texte", "role": "texte", "chars": 1, "index": "texte", "model": "texte", "tokens": "texte", "warnings": "texte", "breakdown": {}, "per_message": "texte", "total_tokens": "texte", "message_count": 1, "context_window": "texte"}}

• conversation_analyze≈ 257 tokens

  Analyze a multi-turn conversation for context retention, topic drift, instruction following, and repetition. Accepts messages array [{role, content}]. Essential for chatbot QA.

  Paramètres, retour & exemple (1)

  Paramètres

  • messages · array · requis
    Conversation messages in order

  Exemple d'appel

  {"name": "conversation_analyze", "arguments": {"messages": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"turn_count": 1, "repetitions": "texte", "topic_drift": "texte", "user_messages": 1, "context_retention": "texte", "has_system_prompt": true, "assistant_messages": 1, "avg_response_length": 1, "repetition_detected": true}}

• cookie_security_audit≈ 324 tokens

  Audit the security attributes of cookies set by any URL. Fetches the URL and inspects all Set-Cookie headers for: HttpOnly, Secure, SameSite, Domain scope, Path scope, Max-Age/Expires, __Host-/__Secure- prefixes. Flags insecure patterns: missing HttpOnly on session cookies, missing Secure flag, SameSite=None without Secure, overly broad Domain, and excessive TTL. Returns per-cookie grades and an overall security score (0–100) — the score is the WEAKEST cookie, not the average, so one leaking session cookie cannot be averaged into a green result (average_score is reported separately).

  Paramètres, retour & exemple (1)

  Paramètres

  • url · string · requis
    Full URL to audit (e.g. https://example.com/login)

  Exemple d'appel

  {"name": "cookie_security_audit", "arguments": {"url": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"url": "texte", "name": "texte", "path": "texte", "score": 1, "domain": "texte", "issues": "texte", "secure": "texte", "cookies": [], "max_age": "texte", "message": "texte", "httpOnly": "texte", "sameSite": "texte"}}

• cors_checker≈ 242 tokens

  Check the CORS configuration of a URL the same way a browser would. Returns the main response status, all Access-Control-* headers, the tested origin, and the preflight OPTIONS response. Use this for direct CORS debugging, not just security auditing.

  Paramètres, retour & exemple (3)

  Paramètres

  • url · string · requis
    Full URL to test, e.g. https://api.example.com/resource
  • method · string · optionnel
    HTTP method to simulate (default: GET)
  • origin · string · optionnel
    Origin header to simulate (default: https://yourdomain.com)

  Exemple d'appel

  {"name": "cors_checker", "arguments": {"url": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"url": "texte", "method": "texte", "status": "texte", "preflight": "texte", "allHeaders": "texte", "corsHeaders": "texte", "testedOrigin": "texte"}}

• cors_test≈ 376 tokens

  Test a URL for CORS misconfigurations. Sends preflight (OPTIONS) requests with various Origin headers to detect: wildcard origins with credentials, origin reflection (echoing any origin), null origin acceptance, subdomain wildcard bypass, and missing Vary headers. Returns risk level (safe/low/medium/high/critical) plus per-origin results. "unknown" means nothing was actually tested — every origin either failed to connect or answered 5xx, so the target returned no CORS decision; never read it as "safe". A 4xx preflight IS a real result (the server refused it and a browser would fail closed).

  Paramètres, retour & exemple (2)

  Paramètres

  • url · string · requis
    Full URL to test (e.g. https://api.example.com/endpoint)
  • origin · string · optionnel
    Custom Origin header to test (default: tests multiple origins automatically)

  Exemple d'appel

  {"name": "cors_test", "arguments": {"url": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"url": "texte", "error": "texte", "tests": "texte", "warning": "texte", "risk_level": "safe", "origins_tested": 1, "total_findings": 1, "origins_reachable": 1, "origins_conclusive": 1, "origins_inconclusive": 1}}

• cot_analyzer≈ 259 tokens

  Analyze a Chain-of-Thought (CoT) or reasoning trace from an LLM. Detects step count, logical flow, conclusion presence, backtracking, and estimates reasoning depth. Useful for o1/o3/DeepSeek-R1 evaluation.

  Paramètres, retour & exemple (2)

  Paramètres

  • reasoning · string · requis
    The CoT / reasoning trace text (e.g. from <think> tags or step-by-step output)
  • expected_conclusion · string · optionnel
    Expected final answer to check against (optional)

  Exemple d'appel

  {"name": "cot_analyzer", "arguments": {"reasoning": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"markers": "texte", "step_count": 1, "total_chars": 1, "total_lines": 1, "has_conclusion": true, "reasoning_depth": "texte", "backtracking_signals": "texte", "reasoning_depth_label": "texte", "conclusion_matches_expected": "texte"}}

• count_code_lines≈ 220 tokens

  Count lines of code: total, code lines, comment lines, blank lines, and comment density. Supports JS/TS, Python, Java/C/C++, Ruby, Go, Shell, HTML/XML, and CSS.

  Paramètres, retour & exemple (2)

  Paramètres

  • input · string · requis
    Source code to analyze
  • language · string · optionnel
    Language hint: "js", "ts", "py", "java", "c", "rb", "go", "sh", "html", "css" (auto-detect if omitted)

  Exemple d'appel

  {"name": "count_code_lines", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"language": "texte", "code_lines": "texte", "blank_lines": "texte", "total_lines": "texte", "comment_lines": "texte", "comment_density": "texte", "code_to_comment_ratio": "texte"}}

• count_tokens≈ 177 tokens

  Estimate the token count of a text string using the cl100k_base approximation (~4 chars/token). Call this BEFORE sending any text to an LLM API to check if it fits within the model context window and to estimate cost. Returns token estimate, character count, and word count.

  Paramètres, retour & exemple (1)

  Paramètres

  • input · string · requis
    Text to count tokens for

  Exemple d'appel

  {"name": "count_tokens", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"chars": "texte", "words": "texte", "tokens_estimate": "texte"}}

• create_confluence_page≈ 428 tokens

  Create a new Confluence page from the output of jira_to_test_suite. Formats Gherkin, E2E steps, API tests, and test data as a properly structured Confluence page with code blocks and tables. STATEFUL — creates a new page in the specified space.

  Paramètres, retour & exemple (9)

  Paramètres

  • title · string · optionnel
    Page title. Defaults to "Test Plan: {issue_key}"
  • issue_key · string · optionnel
    Source Jira issue key (for the page title and source link)
  • issue_url · string · optionnel
    Source Jira issue URL (added as a link in the page)
  • space_key · string · requis
    Confluence space key where the page will be created, e.g. "QA", "ENG"
  • test_suite · object · requis
    The test_suite object from jira_to_test_suite result
  • parent_page_id · string · optionnel
    Optional parent page ID — page will be created as a child of this page
  • confluence_email · string · requis
    Atlassian account email
  • confluence_token · string · requis
    Atlassian API token
  • confluence_base_url · string · requis
    Atlassian base URL

  Exemple d'appel

  {"name": "create_confluence_page", "arguments": {"space_key": "texte", "test_suite": {}, "confluence_email": "texte", "confluence_token": "texte", "confluence_base_url": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"title": "texte", "page_id": "texte", "success": true, "page_url": "texte"}}

• cron_parse≈ 160 tokens

  Parse a cron expression into a human-readable schedule description. Supports standard 5-field cron (minute hour day month weekday).

  Paramètres, retour & exemple (1)

  Paramètres

  • expression · string · requis
    Cron expression (e.g., "0 9 * * 1-5", "*/15 * * * *")

  Exemple d'appel

  {"name": "cron_parse", "arguments": {"expression": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"fields": {}, "expression": "texte", "human_readable": "texte"}}

• cron_validator≈ 223 tokens

  Validate a 5-field cron expression, explain the schedule, and preview the next execution times. Use this to debug cron jobs before they reach production. Returns parsed fields, a human-readable description, and upcoming ISO timestamps.

  Paramètres, retour & exemple (2)

  Paramètres

  • expression · string · requis
    Cron expression with 5 fields, e.g. "*/15 9-18 * * 1-5"
  • next_runs_count · number · optionnel
    How many upcoming runs to return (1-50, default: 10)

  Exemple d'appel

  {"name": "cron_validator", "arguments": {"expression": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"valid": true, "fields": {}, "next_runs": "texte", "expression": "texte", "human_readable": "texte"}}

• decode_jwt≈ 192 tokens

  Decode a JWT (JSON Web Token) and return its header and payload without verifying the signature. Also reports whether the token is expired and the exact expiry date. Use to inspect claims (sub, iss, exp, roles) during debugging or when integrating with an auth provider.

  Paramètres, retour & exemple (1)

  Paramètres

  • token · string · requis
    The JWT string to decode (header.payload.signature)

  Exemple d'appel

  {"name": "decode_jwt", "arguments": {"token": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"note": "texte", "header": "texte", "expired": "texte", "payload": "texte", "expiresAt": "texte"}}

• detect_language≈ 217 tokens

  Detect the natural language of a text using n-gram frequency analysis and common word markers. Supports 15 languages: English, French, Spanish, German, Italian, Portuguese, Dutch, Russian, Chinese, Japanese, Korean, Arabic, Polish, Turkish, Swedish.

  Paramètres, retour & exemple (1)

  Paramètres

  • input · string · requis
    Text to detect language from (min 20 chars for accuracy)

  Exemple d'appel

  {"name": "detect_language", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"lang": "texte", "name": "texte", "score": "texte", "method": "texte", "matched": "texte", "language": "texte", "confidence": 1, "top_candidates": []}}

• detect_secrets≈ 214 tokens

  Scan code or config files for hardcoded secrets: AWS keys, GitHub tokens, OpenAI/Anthropic API keys, Stripe secrets, JWTs, database connection strings, and generic passwords. Returns findings with severity. Run before every commit.

  Paramètres, retour & exemple (2)

  Paramètres

  • input · string · requis
    Code or config content to scan (max 500KB)
  • filename · string · optionnel
    Optional filename for context (e.g. ".env", "config.js")

  Exemple d'appel

  {"name": "detect_secrets", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"filename": "texte", "findings": "texte", "risk_level": "texte", "recommendation": "texte", "total_findings": 1}}

• diff_mappings≈ 315 tokens

  Diff a baseline page mapping against a current one and return a CI-style verdict: PASS / FIX / BLOCK, plus per-element drift (ok, renamed, healable, ambiguous, lost, added, rebound). Pure and deterministic — provide two mappings as JSON with "elements" arrays of {role, name, selector, context?}. Use the companion @ia-qa/self-healing package (npm install -g @ia-qa/self-healing) to capture mappings from your app via its local MCP server ia-qa-heal-mcp, or paste the snippet from ia-qa.com/devtools/selector-drift into your browser console.

  Paramètres, retour & exemple (2)

  Paramètres

  • after · object · requis
    Current page mapping: same shape as before, captured after the UI change.
  • before · object · requis
    Baseline page mapping: { page, url, capturedAt, elements: [{role, name, selector, context?}] }. Captured before a UI change.

  Exemple d'appel

  {"name": "diff_mappings", "arguments": {"after": {}, "before": {}}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"rows": [], "added": [], "counts": {}, "verdict": "texte"}}

• diff_text≈ 241 tokens

  Compute a unified line-by-line diff between two text strings (LCS algorithm). Returns added/removed/unchanged line counts and formatted diff hunks with configurable context lines (0–20). Use to compare versions of prompts, configs, code snippets, or any text where you need to see exactly what changed.

  Paramètres, retour & exemple (3)

  Paramètres

  • a · string · requis
    Original (before) text
  • b · string · requis
    Modified (after) text
  • context · number · optionnel
    Context lines around each change (0–20, default: 3)

  Exemple d'appel

  {"name": "diff_text", "arguments": {"a": "texte", "b": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"diff": "texte", "added": 1, "removed": 1, "unchanged": 1}}

• embedding_similarity≈ 370 tokens

  Compute text similarity using local algorithms (Bag of Words, TF-IDF, Character N-grams). No API key needed — runs entirely in-process. NOT real embeddings: for true semantic similarity with vector embeddings, use run_semantic_tests with mode="embeddings" and your OpenAI API key. Supports single pair or batch mode with pipe-separated pairs. Useful for RAG retrieval testing, semantic search evaluation, and text deduplication.

  Paramètres, retour & exemple (4)

  Paramètres

  • batch · array · optionnel
    Batch mode: array of { text_a, text_b } pairs. Overrides text_a/text_b if provided.
  • text_a · string · optionnel
    First text to compare (single-pair mode)
  • text_b · string · optionnel
    Second text to compare (single-pair mode)
  • methods · array · optionnel
    Algorithms to use (default: all three). Options: "bow", "tfidf", "ngram"

  Exemple d'appel

  {"name": "embedding_similarity", "arguments": {"batch": [], "text_a": "texte", "text_b": "texte", "methods": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"mode": "texte", "count": 1, "scores": "texte", "text_a": "texte", "text_b": "texte", "results": "texte"}}

• env_parse≈ 189 tokens

  Parse a .env file content into a JSON object. Handles quoted values (single and double), inline comments, export prefix, and escaped sequences (\n, \t inside double quotes). Returns all key-value pairs. Use in CI/CD pipelines, agent config loaders, or when processing dotenv files programmatically.

  Paramètres, retour & exemple (1)

  Paramètres

  • input · string · requis
    .env file content to parse (e.g. the output of `cat .env`)

  Exemple d'appel

  {"name": "env_parse", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"vars": "texte", "count": 1}}

• escape_html≈ 164 tokens

  Escape HTML special characters (&, <, >, ", ') to their safe HTML entities. ALWAYS call this before inserting any user-provided or LLM-generated content into an HTML template to prevent cross-site scripting (XSS) attacks.

  Paramètres, retour & exemple (1)

  Paramètres

  • input · string · requis
    String to HTML-escape

  Exemple d'appel

  {"name": "escape_html", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"escaped": "texte", "original_length": "texte"}}

• estimate_llm_cost≈ 266 tokens

  Estimate the API cost in USD for a given model and token counts. Supports all major 2024–2026 models: GPT-4o, GPT-4.1, o3, o4-mini, Claude Opus 4, Claude Sonnet 4/4.5, Gemini 2.5 Pro/Flash, DeepSeek V3/R1, Grok 3, and legacy models.

  Paramètres, retour & exemple (3)

  Paramètres

  • model · string · requis
    Model name, e.g. "gpt-4o", "claude-3.5-sonnet", "deepseek-v3"
  • input_tokens · number · requis
    Number of input/prompt tokens
  • output_tokens · number · optionnel
    Number of output/completion tokens (default: 0)

  Exemple d'appel

  {"name": "estimate_llm_cost", "arguments": {"model": "texte", "input_tokens": 1}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"model": "texte", "rates": {}, "input_tokens": "texte", "output_tokens": "texte", "input_cost_usd": "texte", "total_cost_usd": "texte", "output_cost_usd": "texte"}}

• extract_json_from_text≈ 188 tokens

  Extract the first valid JSON object or array embedded in chaotic LLM output (surrounded by markdown fences, prose, or explanatory text). Handles ```json blocks and inline JSON. Call this whenever an LLM returns structured data mixed with explanation text instead of raw JSON.

  Paramètres, retour & exemple (1)

  Paramètres

  • input · string · requis
    Raw text (e.g., LLM output) that may contain a JSON object or array

  Exemple d'appel

  {"name": "extract_json_from_text", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"json": "texte", "source": "texte"}}

• extract_json_path≈ 196 tokens

  Extract a value from a JSON string using dot-notation path (e.g., "user.address.city", "items.0.name", "meta.tags"). Supports array index access via numeric path segments.

  Paramètres, retour & exemple (2)

  Paramètres

  • path · string · requis
    Dot-notation path, e.g. "user.address.city" or "items.0.name"
  • input · string|object|array · requis
    The JSON to traverse — a JSON string, or the object/array itself.

  Exemple d'appel

  {"name": "extract_json_path", "arguments": {"path": "texte", "input": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"path": "texte", "type": "texte", "value": "texte"}}

• extract_links≈ 182 tokens

  Extract all URLs, email addresses, and domain names from text. Returns categorized and deduplicated results. Useful for content auditing, link checking, and web scraping validation.

  Paramètres, retour & exemple (2)

  Paramètres

  • input · string · requis
    Text to extract links from
  • types · array · optionnel
    Types to extract (default: all three)

  Exemple d'appel

  {"name": "extract_links", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"total": "texte"}}

• extract_todos≈ 226 tokens

  Extract TODO, FIXME, HACK, BUG, NOTE, OPTIMIZE, and custom tags from any source code or text. Returns line numbers, tag types, and message text. Essential for technical debt auditing.

  Paramètres, retour & exemple (3)

  Paramètres

  • tags · array · optionnel
    Custom tags to add (default set: TODO, FIXME, HACK, NOTE, BUG, OPTIMIZE, XXX)
  • input · string · requis
    Code or text to scan
  • include_context · boolean · optionnel
    Include full line text (default: true)

  Exemple d'appel

  {"name": "extract_todos", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"items": "texte", "total": 1, "counts": "texte", "has_critical": true}}

• fetch_confluence_page≈ 365 tokens

  Fetch a Confluence page and return its content as clean Markdown. Accepts a numeric page_id or a full page URL. Optionally lists direct child pages. BYOK — credentials transit in-memory only, never stored.

  Paramètres, retour & exemple (6)

  Paramètres

  • page_id · string · optionnel
    Confluence page ID (numeric string), e.g. "123456789"
  • page_url · string · optionnel
    Full Confluence page URL (alternative to page_id), e.g. "https://mycompany.atlassian.net/wiki/spaces/ENG/pages/123456789"
  • confluence_email · string · requis
    Atlassian account email (same credentials as Jira)
  • confluence_token · string · requis
    Atlassian API token
  • include_children · boolean · optionnel
    List direct child pages (id + title) (default: false)
  • confluence_base_url · string · requis
    Atlassian base URL, e.g. "https://mycompany.atlassian.net"

  Exemple d'appel

  {"name": "fetch_confluence_page", "arguments": {"confluence_email": "texte", "confluence_token": "texte", "confluence_base_url": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"url": "texte", "title": "texte", "page_id": "texte", "children": [], "markdown": "texte"}}

• fetch_jira_issue≈ 425 tokens

  Fetch a complete Jira issue: summary, description converted to Markdown, status, assignee, priority, labels, custom fields, and optionally comments and attachment metadata. BYOK — credentials transit in-memory only, never stored on ia-qa.com.

  Paramètres, retour & exemple (7)

  Paramètres

  • fields · array · optionnel
    Specific Jira field names to return. Omit for all standard fields.
  • issue_key · string · requis
    Jira issue key, e.g. "PROJ-123"
  • jira_email · string · requis
    Atlassian account email
  • jira_token · string · requis
    Atlassian API token (from id.atlassian.com > Security > API tokens)
  • jira_base_url · string · requis
    Atlassian base URL, e.g. "https://mycompany.atlassian.net"
  • include_comments · boolean · optionnel
    Include issue comments, up to 20 (default: true)
  • include_attachments · boolean · optionnel
    Include attachment metadata list (default: false)

  Exemple d'appel

  {"name": "fetch_jira_issue", "arguments": {"issue_key": "texte", "jira_email": "texte", "jira_token": "texte", "jira_base_url": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"key": "texte", "url": "texte", "type": "texte", "labels": [], "status": "texte", "summary": "texte", "assignee": "texte", "priority": "texte", "reporter": "texte", "description": "texte"}}

• fetch_veille_feed≈ 272 tokens

  Fetch the latest QA & AI/LLM articles aggregated from curated RSS sources (Google Testing Blog, DEV.to Testing/QA/AI/LLM/Agents, Hugging Face Blog, Simon Willison). Perfect for agents monitoring the QA & AI landscape. Each article carries summary_source — the XML tag the summary was read from, or "none" when the feed publishes titles and links only; an empty summary with summary_source "none" is a property of that feed, not a parse failure.

  Paramètres, retour & exemple (2)

  Paramètres

  • limit · number · optionnel
    Max articles to return (default: 20, max: 50)
  • category · string · optionnel
    Filter: "qa" (testing/quality), "ai" (AI/LLM/agents), "all" (default — both)

  Exemple d'appel

  {"name": "fetch_veille_feed", "arguments": {"limit": 1, "category": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"articles": [], "category": "texte", "total_found": 1, "sources_queried": 1}}

• few_shot_formatter≈ 281 tokens

  Format few-shot examples for LLM prompts. Converts example pairs into formatted blocks. Supports chat format (User/Assistant), XML tags, Markdown, or plain text.

  Paramètres, retour & exemple (4)

  Paramètres

  • format · string · optionnel · chat | xml | markdown | plain
    Output format (default: chat)
  • examples · array · requis
    Array of {input, output} pairs
  • input_label · string · optionnel
    Label for input (default: User / <input>)
  • output_label · string · optionnel
    Label for output (default: Assistant / <output>)

  Exemple d'appel

  {"name": "few_shot_formatter", "arguments": {"examples": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"format": "texte", "formatted": "texte", "example_count": 1, "token_estimate": 1}}

• find_tool≈ 351 tokens

  Search available MCP tools by keyword or category before calling them. Returns matching tool names, descriptions, and optionally their inputSchemas. Call this when you are unsure which tool to use or want to explore the catalogue. Categories: data, encoding, text, llm, qa, rag, dev, security, web.

  Paramètres, retour & exemple (4)

  Paramètres

  • query · string · requis
    Keyword(s) to search in tool name and description (e.g. "cors", "token", "vector", "json")
  • category · string · optionnel
    Optional: filter by category — data | encoding | text | llm | qa | rag | dev | security | web
  • max_results · number · optionnel
    Maximum tools to return (default 10, max 50). Results are ranked by IDF-weighted relevance, so common words like "test" do not inflate the list.
  • with_schema · boolean · optionnel
    Set true to include inputSchema in results (default: false)

  Exemple d'appel

  {"name": "find_tool", "arguments": {"query": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"hint": "texte", "tool": "texte", "count": 1, "query": "texte", "score": "texte", "tools": [], "category": "texte", "truncated": true, "total_matches": 1}}

• fix_gherkin≈ 437 tokens

  Fix Gherkin syntax warnings from a jira_to_test_suite result. Takes the current gherkin text and the _gherkin_warnings array, calls your LLM to fix ONLY the flagged issues (adds missing Given/When/Then steps, etc.), and returns the corrected Gherkin. Lightweight — uses ~300-500 tokens vs ~5k for a full regeneration. Requires BYOK LLM key.

  Paramètres, retour & exemple (4)

  Paramètres

  • model · string · requis
    LLM model to use for the fix, e.g. "gpt-4o-mini". Must belong to the provider whose key you passed in api_key.
  • api_key · string · requis
    Your own LLM provider API key (BYOK) — OpenAI "sk-…", Anthropic "sk-ant-…", Google "AIzaSy…", or Groq "gsk_…". There is no server-side key for this tool: if you do not have one, do not call it and do not invent a value — placeholders like "configured", "your_api_key" or a masked "sk-…***…" are rejected. Used for this call only, never stored.
  • gherkin · string · requis
    The current Gherkin text from the jira_to_test_suite result (test_suite.gherkin).
  • warnings · array · requis
    The _gherkin_warnings array from the jira_to_test_suite result.

  Exemple d'appel

  {"name": "fix_gherkin", "arguments": {"model": "texte", "api_key": "texte", "gherkin": "texte", "warnings": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"latency_ms": 1, "model_used": "texte", "fixed_gherkin": "texte", "warnings_after": 1, "warnings_before": 1, "remaining_warnings": []}}

• flatten_json≈ 221 tokens

  Flatten a nested JSON object to single-level dot-notation keys (e.g. {"a":{"b":1}} → {"a.b":1}), or unflatten dot-notation keys back to a nested object. Supports custom separators.

  Paramètres, retour & exemple (3)

  Paramètres

  • mode · string · optionnel
    "flatten" (default) or "unflatten"
  • input · string|object|array · requis
    The JSON to flatten or unflatten — a JSON string, or the object itself.
  • separator · string · optionnel
    Key separator (default: ".")

  Exemple d'appel

  {"name": "flatten_json", "arguments": {"input": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"result": "texte", "key_count": 1, "max_depth": []}}

• format_bytes≈ 202 tokens

  Convert raw byte counts to human-readable sizes in SI (KB=1000) or IEC (KiB=1024) units, or parse size strings back to bytes. Covers B, KB/KiB, MB/MiB, GB/GiB, TB/TiB, PB/PiB.

  Paramètres, retour & exemple (3)

  Paramètres

  • bytes · number · optionnel
    Number of bytes to format
  • standard · string · optionnel · both | si | iec
    Output standard (default: both)
  • size_string · string · optionnel
    Size string to parse to bytes (e.g. "1.5 GB", "512 MiB")

  Exemple d'appel

  {"name": "format_bytes", "arguments": {"bytes": 1, "standard": "both", "size_string": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"bytes": 1, "original": "texte"}}

• format_json≈ 258 tokens

  Validate and pretty-print a string that is ALREADY valid JSON. Strict by design — it is a validity gate: valid JSON comes back formatted, anything else is rejected with the exact parse error. It never repairs, completes, or guesses. NOT for: plain text or prose (will fail), JSON embedded in markdown/prose (use extract_json_from_text first), JS objects (JSON.stringify them first), YAML (use yaml_to_json).

  Paramètres, retour & exemple (2)

  Paramètres

  • input · string · requis
    A raw JSON string, e.g. '{"key":"value"}'. Must already parse as JSON — plain text or truncated JSON is rejected, not repaired.
  • indent · number · optionnel
    Indent size (default: 2)

  Exemple d'appel

  {"name": "format_json", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"valid": true, "formatted": "texte"}}

• format_table≈ 240 tokens

  Convert a JSON array of objects into a Markdown table. Automatically detects columns, aligns headers, and fills missing keys with empty cells. Use when an agent needs to present structured data — tool results, model comparisons, test reports — as a readable table in a response or document.

  Paramètres, retour & exemple (2)

  Paramètres

  • input · string|object|array · requis
    The array of objects to render — a JSON string, or the array itself.
  • columns · array · optionnel
    Column names and order (default: all keys from first row)

  Exemple d'appel

  {"name": "format_table", "arguments": {"input": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"rows": 1, "table": "texte", "columns": 1}}

• function_call_validate≈ 300 tokens

  Validate an LLM function call / tool_use output: check that function name is in allowed list, arguments match expected schema, no extra/missing args. For OpenAI function calling & MCP tool_use testing.

  Paramètres, retour & exemple (2)

  Paramètres

  • function_call · object · requis
    The function call object from LLM (e.g. { "name": "get_weather", "arguments": {"city":"Paris"} })
  • allowed_functions · array · requis
    List of allowed function definitions

  Exemple d'appel

  {"name": "function_call_validate", "arguments": {"function_call": {}, "allowed_functions": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"valid": true, "errors": [], "error_count": 1, "function_name": "texte", "provided_args": "texte", "required_args": "texte"}}

• generate_ci_workflow≈ 898 tokens

  Generate a ready-to-commit GitHub Actions workflow that gates a build on IA-QA. Two gate types, combinable: "eval_contract" runs a .ia-eval.yaml through ia-qa-com/eval-action@v1 (LLM quality gate, needs a provider API key as a repo secret), and "cli_checks" runs deterministic primitives via npx @ia-qa/cli (secret scan, prompt-injection scan, security headers…) whose exit code fails the build. Deterministic template — no LLM call, no API key, same inputs give the same file. Returns the YAML, the secrets to create, and the remaining steps. Pair with generate_eval_yaml to produce the contract itself.

  Paramètres, retour & exemple (12)

  Paramètres

  • cron · string · optionnel
    Cron expression when triggers include 'schedule' (default: '0 6 * * 1' — Mondays 06:00 UTC).
  • gate · string · optionnel · eval_contract | cli_checks | both | selector_drift | all
    Which gate to emit. eval_contract = LLM eval via the action (default). cli_checks = deterministic CLI assertions. selector_drift = an E2E selector-drift gate via @ia-qa/self-healing (boots the app, captures, diffs against the committed baseline, branches on exit code 0/1/2). both = CLI checks + eval. all = CLI checks, then drift, then eval.
  • provider · string · optionnel · groq | openai | anthropic | google
    LLM provider the contract runs against — decides which repository secret the workflow wires (default: groq).
  • triggers · array · optionnel
    Workflow triggers (default: push + pull_request).
  • cli_tools · array · optionnel
    IA-QA tool names to run as deterministic gates, e.g. ["secret_scan","prompt_injection_scan"]. Tools with no known CI recipe get a --stdin step flagged in notes.
  • min_score · number · optionnel
    Override the contract min_score (0-100). Omit to use the value in the contract.
  • app_base_url · string · optionnel
    URL the drift gate waits for before capturing (default: http://127.0.0.1:3000). Must match config.baseUrl in .ia-qa/config.json.
  • fail_on_fail · boolean · optionnel
    Fail the build on a FAIL/PARTIAL verdict (default: true). Set false to report without gating.
  • node_version · string · optionnel
    Node version for the CLI steps (default: "20").
  • contract_path · string · optionnel
    Path to the .ia-eval.yaml contract, relative to the repo root (default: evals/smoke.ia-eval.yaml). Only used when the gate includes eval_contract.
  • workflow_name · string · optionnel
    Workflow display name (default: "IA-QA Quality Gate").
  • app_start_command · string · optionnel
    Command that boots the app for the selector_drift gate, e.g. "npm run start:ci". Cannot be guessed — omitted, the step carries an explicit TODO and the note says so, because a plausible default would silently map nothing.

  Exemple d'appel

  {"name": "generate_ci_workflow", "arguments": {"cron": "texte", "gate": "eval_contract", "provider": "groq", "triggers": [], "cli_tools": [], "min_score": 1}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"gate": "texte", "path": "texte", "yaml": "texte", "notes": [], "next_steps": [], "secrets_required": []}}

• generate_curl≈ 312 tokens

  Generate a curl command from request parameters. Supports GET/POST/PUT/DELETE, custom headers, JSON body, and form data. Useful for documentation, sharing, and debugging API calls.

  Paramètres, retour & exemple (7)

  Paramètres

  • url · string · requis
    Request URL (must be http/https)
  • body · string · optionnel
    Raw request body string
  • method · string · optionnel
    HTTP method (default: GET)
  • headers · object · optionnel
    Request headers as key-value object
  • verbose · boolean · optionnel
    Add -v for verbose output (default: false)
  • body_json · object · optionnel
    JSON body (auto-adds Content-Type: application/json)
  • follow_redirects · boolean · optionnel
    Follow redirects with -L flag (default: true)

  Exemple d'appel

  {"name": "generate_curl", "arguments": {"url": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"url": "texte", "curl": "texte", "method": "texte", "header_count": 1}}

• generate_eval_yaml≈ 379 tokens

  Generate a complete .ia-eval.yaml evaluation contract from a plain-language description of what your LLM should do. Uses Groq openai/gpt-oss-20b (server-side, no API key needed). Returns ready-to-run YAML for the LLM Test Runner (run_eval_contract). Picks appropriate evaluators (cosine_similarity, contains_check, hallucination_check, etc.) based on the task type.

  Paramètres, retour & exemple (4)

  Paramètres

  • task_type · string · optionnel · rag | summarization | classification | safety | customer_support | code_gen
    Optional task type hint to guide evaluator selection.
  • description · string · requis
    Plain-language description of what the LLM under test should do. Be specific: describe inputs, expected behaviour, and constraints.
  • system_prompt · string · optionnel
    Optional system prompt of the LLM under test. Helps generate more accurate test cases.
  • scenario_count · number · optionnel · 3 | 5 | 8
    Number of scenarios to generate (default: 5). Covers happy path + edge cases + adversarial.

  Exemple d'appel

  {"name": "generate_eval_yaml", "arguments": {"description": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"yaml": "texte", "task_type": "texte", "model_used": "texte", "scenario_count": 1}}

• generate_hmac≈ 241 tokens

  Compute an HMAC signature for a message using a secret key. Supports SHA-256 (default), SHA-512, SHA-1, and MD5. Used for API request signing, webhook verification (GitHub, Stripe, Twilio), and JWT validation.

  Paramètres, retour & exemple (4)

  Paramètres

  • secret · string · requis
    Secret key
  • message · string · requis
    Message to sign
  • encoding · string · optionnel · hex | base64 | base64url
    Output encoding (default: hex)
  • algorithm · string · optionnel
    Hash algorithm: sha256 (default), sha512, sha1, md5

  Exemple d'appel

  {"name": "generate_hmac", "arguments": {"secret": "texte", "message": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"hmac": "texte", "encoding": "texte", "algorithm": "texte", "message_length": 1}}

• generate_html_report≈ 178 tokens

  Convert a run_eval_contract() LLM Test Runner JSON result into a fully self-contained dark-themed HTML report with Pass/Fail badges, side-by-side Input/Output/Ground-Truth panels, evaluator score bars, and a radar chart. Returns the HTML as a string.

  Paramètres, retour & exemple (1)

  Paramètres

  • results · object · requis
    The JSON object returned by run_eval_contract()

  Exemple d'appel

  {"name": "generate_html_report", "arguments": {"results": {}}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"html": "texte"}}

• generate_json_ld≈ 294 tokens

  Generate a ready-to-paste <script type="application/ld+json"> snippet for GEO / structured data optimization. Supported types: WebSite, FAQPage, Article, Person, Organization, SoftwareApplication, HowTo.

  Paramètres, retour & exemple (3)

  Paramètres

  • type · string · requis
    Schema @type: "WebSite", "FAQPage", "Article", "Person", "Organization", "SoftwareApplication", "HowTo"
  • fields · object · optionnel
    Schema fields as key-value pairs (name, url, description, author, datePublished, etc.)
  • faq_items · array · optionnel
    For FAQPage/HowTo: array of { question, answer } objects

  Exemple d'appel

  {"name": "generate_json_ld", "arguments": {"type": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"name": "texte", "schema": "texte", "snippet": "texte", "acceptedAnswer": {}}}

• generate_password≈ 240 tokens

  Generate a cryptographically secure random password using crypto.randomBytes. Configurable length (4–128), uppercase letters, digits, and symbols. Use when resetting user passwords, seeding test accounts, or generating API secrets.

  Paramètres, retour & exemple (4)

  Paramètres

  • length · number · optionnel
    Password length (4–128, default: 16)
  • numbers · boolean · optionnel
    Include digits (default: true)
  • symbols · boolean · optionnel
    Include symbols like !@#$ (default: false)
  • uppercase · boolean · optionnel
    Include uppercase letters (default: true)

  Exemple d'appel

  {"name": "generate_password", "arguments": {"length": 1, "numbers": true, "symbols": true, "uppercase": true}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"length": 1, "password": "texte", "charset_size": 1}}

• generate_slug≈ 182 tokens

  Convert any string into a URL-friendly slug: lowercase, ASCII-normalized (é→e), special characters removed, spaces replaced with hyphens. Use for generating SEO-friendly URL paths, file names, or identifier keys from user-provided titles or labels.

  Paramètres, retour & exemple (2)

  Paramètres

  • input · string · requis
    String to slugify
  • separator · string · optionnel
    Separator character (default: "-")

  Exemple d'appel

  {"name": "generate_slug", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"slug": "texte"}}

• generate_test_cases≈ 427 tokens

  Generate a set of test cases (valid, edge, invalid, pairwise) for a given feature description. Declared constraints drive the boundaries: a length or numeric bound ("[8-64]", "min 8 chars", "at least 8 characters" — read from inputs, and from the feature prose when a sentence names exactly one field) yields the last accepted value AND the first rejected one; a format (email/url/uuid, from the type, the field name or the wording) yields malformed-value cases. A bound nobody declared is labelled as this tool's assumption, not as expected behaviour. gherkinFormat renders every case (cap 200, stated in the output) and gherkinScenarioCount lets you check it against totalCases.

  Paramètres, retour & exemple (2)

  Paramètres

  • inputs · string · optionnel
    Optional: list of input parameters (one per line, e.g. "email: string [required]", "password: string [required, min 8 chars]", "age: number [18-99]")
  • feature · string · requis
    Feature or function to test. Be specific: describe inputs, expected behaviour, context. Constraints stated here ("password must be at least 8 characters") are used when the sentence names exactly one field.

  Exemple d'appel

  {"name": "generate_test_cases", "arguments": {"feature": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"feature": "texte", "test_cases": [], "parsedInputs": [], "gherkinFormat": "texte", "gherkinScenarioCount": 1}}

• generate_uuid≈ 168 tokens

  Generate one or more cryptographically random UUID v4 identifiers. Use this when you need unique IDs for test fixtures, database records, session tokens, or any scenario requiring a guaranteed-unique string. Returns up to 100 UUIDs in one call.

  Paramètres, retour & exemple (1)

  Paramètres

  • count · number · optionnel
    Number of UUIDs to generate (1–100, default: 1)

  Exemple d'appel

  {"name": "generate_uuid", "arguments": {"count": 1}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"count": 1, "uuids": "texte"}}

• get_testing_guidelines≈ 452 tokens

  Query the IA-QA methodology knowledge base. Returns structured testing guidelines, assertion strategies, thresholds, best practices, and relevant MCP tools for a given topic. Call without a topic to list all available topics. Topics: llm-unit-testing, rag-pipeline, prompt-stability, prompt-ab-testing, embedding-quality, eval-framework, semantic-testing, auto-testing, security, api-testing, ci-cd, multimodal, llm-data-security, agent-observability, pro-tips, learning-paths, golden-dataset, selector-drift, qa-recipes, playbooks. Not sure where to start testing an LLM, RAG pipeline or agent? Call without a topic (or with "start-here"): it maps what you are testing to the tools to call and the output field to gate CI on. A plain question such as "how do I test my RAG" also resolves to the right topic.

  Paramètres, retour & exemple (1)

  Paramètres

  • topic · string · optionnel · start-here | llm-unit-testing | rag-pipeline | prompt-stability | prompt-ab-testing | embedding-quality | eval-framework | semantic-testing
    The testing topic to retrieve guidelines for. Omit to get the start-here map and the full list of available topics.

  Exemple d'appel

  {"name": "get_testing_guidelines", "arguments": {"topic": "start-here"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"tip": "texte", "topic": "texte", "usage": "texte", "keywords": [], "start_here": {}, "available_topics": []}}

• guardrail_test≈ 332 tokens

  Test an LLM response against a set of guardrail rules: must-include, must-not-include, max length, required format, language, forbidden patterns, and custom regex. Returns pass/fail per rule.

  Paramètres, retour & exemple (2)

  Paramètres

  • rules · array · requis
    Array of guardrail rules to check
  • response · string · requis
    The LLM response to test

  Exemple d'appel

  {"name": "guardrail_test", "arguments": {"rules": [], "response": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"pass": true, "rule": "texte", "label": "texte", "value": "texte", "detail": "texte", "failed": "texte", "passed": "texte", "results": "texte", "all_passed": true, "total_rules": 1}}

• hallucination_check≈ 480 tokens

  Lexical hallucination check: verifies an LLM answer's words, numbers and polarity against the provided source/context. Fast, deterministic, no API key needed. Each answer sentence is aligned to its best-matching source sentence, so a number only counts as support when it sits on the SAME statement ("founded in 1998" is not grounded by "sold 1998 units"), and a negation or antonym flip against that sentence returns verdict "contradicted" — the corrupted-fact hallucination that reuses source vocabulary. Limitations: still lexical — it cannot follow a paraphrase, a synonym, or multi-sentence reasoning, so a "well_grounded" verdict means "nothing lexical found", never "verified true". For entailment use run_semantic_tests (NLI/embedding) or a calibrated judge.

  Paramètres, retour & exemple (3)

  Paramètres

  • answer · string · requis
    The LLM-generated answer to verify
  • strict · boolean · optionnel
    If true, every sentence in the answer must be supported (default: false)
  • context · string · requis
    The source/reference text that should ground the answer

  Exemple d'appel

  {"name": "hallucination_check", "arguments": {"answer": "texte", "context": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"detail": "texte", "message": "texte", "numbers": "texte", "overlap": 1, "verdict": "texte", "analysis": "texte", "entities": "texte", "grounded": true, "sentence": "texte", "total_words": 1, "matched_words": 1, "contradictions": "texte"}}

• hash_text≈ 198 tokens

  Compute a cryptographic hash of a text string. Use when you need to verify data integrity, generate content fingerprints, hash passwords (prefer SHA-256+), or produce a fixed-length digest of any input. Supports SHA-256 (default), SHA-512, SHA-1, and MD5.

  Paramètres, retour & exemple (2)

  Paramètres

  • input · string · requis
    Text to hash
  • algorithm · string · optionnel
    Hash algorithm: sha256 (default), sha512, sha1, md5

  Exemple d'appel

  {"name": "hash_text", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"hash": "texte", "algorithm": "texte", "input_length": 1}}

• html_to_markdown≈ 191 tokens

  Convert HTML to clean Markdown. Strips scripts, styles, nav, ads, and comments. Converts headings, lists, links, images, code blocks. Ideal for preparing web content as LLM context.

  Paramètres, retour & exemple (2)

  Paramètres

  • input · string · requis
    HTML string to convert
  • strip_links · boolean · optionnel
    Strip link URLs, keep text only (default: false)

  Exemple d'appel

  {"name": "html_to_markdown", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"markdown": "texte", "markdown_length": 1, "original_length": 1}}

• http_status_lookup≈ 251 tokens

  Look up detailed information about any HTTP status code: class, name, description, cacheability, typical causes, and handling best practices. Covers every code in the IANA HTTP Status Code Registry (1xx-5xx, including 226, 425, 451, 508, 511 and the WebDAV codes) with its defining RFC; anything outside the registry is reported as `registered: false` rather than described. `cacheable` means heuristically cacheable by default per RFC 9110 §15.1.

  Paramètres, retour & exemple (1)

  Paramètres

  • code · number · requis
    HTTP status code (e.g. 200, 404, 429, 503)

  Exemple d'appel

  {"name": "http_status_lookup", "arguments": {"code": 1}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"code": "texte", "desc": "texte", "name": "texte", "class": "texte", "cacheable": "texte", "registered": true, "description": "texte"}}

• identify_caller≈ 216 tokens

  Returns what the server knows about the current MCP client: clientInfo captured during initialize, User-Agent, and any _meta fields sent with this request. Useful for debugging caller identification.

  Paramètres, retour & exemple (1)

  Paramètres

  • _meta · object · optionnel
    Optional self-identification. Keys: agent (string), model (string), version (string).

  Exemple d'appel

  {"name": "identify_caller", "arguments": {"_meta": {}}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"note": "texte", "session": {}, "meta_override": {}, "effective_agent": "texte"}}

• jira_to_test_suite≈ 602 tokens

  Transform a Jira ticket into a complete test suite: Gherkin scenarios, E2E steps, API test cases, test data matrix, and ambiguity detection. Accepts either Jira credentials (auto-fetch) or a pre-fetched issue object. The returned test_suite includes _gherkin_warnings (deterministic syntax validation — empty if clean). Requires BYOK LLM key (OpenAI, Anthropic, etc.).

  Paramètres, retour & exemple (9)

  Paramètres

  • issue · object · optionnel
    Pre-fetched issue object from fetch_jira_issue, OR a mock object with fields: key, summary, description (plain text or Markdown), status, issue_type, priority, labels, comments. Use this for offline/CI testing without Jira credentials.
  • model · string · requis
    LLM model to use, e.g. "gpt-4o-mini", "claude-3-5-haiku-20241022", "gemini-2.0-flash".
  • api_key · string · requis
    Your LLM provider API key (OpenAI sk-, Anthropic sk-ant-, Google AIzaSy-, etc.).
  • issue_key · string · optionnel
    Jira issue key to fetch automatically, e.g. "PROJ-123". Required if issue is not provided.
  • jira_email · string · optionnel
    Atlassian account email. Required for auto-fetch mode.
  • jira_token · string · optionnel
    Atlassian API token. Required for auto-fetch mode.
  • max_tokens · integer · optionnel
    Maximum tokens for the LLM response. Default: 8192. Increase for large tickets with many ACs; decrease to reduce cost on simple tickets.
  • jira_base_url · string · optionnel
    Atlassian base URL. Required for auto-fetch mode.
  • confluence_pages · array · optionnel
    Optional array of pre-fetched Confluence page objects from fetch_confluence_page, used as documentation context.

  Exemple d'appel

  {"name": "jira_to_test_suite", "arguments": {"model": "texte", "api_key": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"summary": "texte", "issue_key": "texte", "issue_url": "texte", "latency_ms": 1, "model_used": "texte", "test_suite": {}, "tokens_used": 1}}

• json_diff≈ 346 tokens

  Compute a deep structural diff between two JSON values. Returns added, removed, and changed keys with dot-notation paths. Like git diff but for JSON objects — perfect for API response regression testing. Arrays are not compared blindly by position: the same elements in a different order collapse to a single "reordered" change, and an array of records sharing a stable identity field (id, uuid, key, name…) is matched by that field, so paths read [id=42] and a moved record is not reported as N rewrites.

  Paramètres, retour & exemple (3)

  Paramètres

  • after · string|object|array · requis
    The modified JSON (after) — a JSON string, or the value itself.
  • before · string|object|array · requis
    The original JSON (before) — a JSON string, or the value itself.
  • max_depth · number · optionnel
    Max nesting depth to recurse (default: 10)

  Exemple d'appel

  {"name": "json_diff", "arguments": {"after": [], "before": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"added": true, "changes": [], "removed": true, "modified": true, "identical": true, "total_changes": 1}}

• json_schema_generate≈ 266 tokens

  Infer a JSON Schema (draft-07) from a sample JSON value. Detects types, required fields, array item shapes, nested objects, and common string formats (email, uri, date, date-time, uuid). Returns a ready-to-use schema compatible with json_schema_validate. Use when you have a sample API response or LLM output and want to auto-generate a validation schema for CI/CD testing.

  Paramètres, retour & exemple (2)

  Paramètres

  • input · string|object|array · requis
    The sample JSON value to infer the schema from — a JSON string, or the value itself.
  • required_all · boolean · optionnel
    Mark all detected object properties as required (default: true)

  Exemple d'appel

  {"name": "json_schema_generate", "arguments": {"input": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"type": "texte", "items": {}, "format": "texte", "schema": {}}}

• json_schema_validate≈ 252 tokens

  Validate a JSON value against a JSON Schema (draft-07 subset). Supports type, required, properties, items, enum, const, pattern, format (email/uri/date), minimum/maximum, minLength/maxLength, minItems/maxItems, uniqueItems, additionalProperties, anyOf, allOf, oneOf. Returns all validation errors with dot-notation paths.

  Paramètres, retour & exemple (2)

  Paramètres

  • value · string|object|array · requis
    The JSON value to validate — a JSON string, or the value itself.
  • schema · string|object|array · requis
    The JSON Schema — a JSON string, or the schema object itself.

  Exemple d'appel

  {"name": "json_schema_validate", "arguments": {"value": [], "schema": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"valid": true, "errors": [], "error_count": 1}}

• json_to_csv≈ 216 tokens

  Convert a JSON array of objects to CSV format. Automatically detects columns from all object keys. Handles quoting and escaping per RFC 4180.

  Paramètres, retour & exemple (3)

  Paramètres

  • input · string|object|array · requis
    The array of objects to convert — a JSON string, or the array itself.
  • headers · boolean · optionnel
    Include header row (default: true)
  • delimiter · string · optionnel
    Column delimiter (default: ",")

  Exemple d'appel

  {"name": "json_to_csv", "arguments": {"input": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"csv": "texte", "rows": 1, "columns": 1, "column_names": "texte"}}

• json_to_yaml≈ 182 tokens

  Convert a JSON object to clean, human-readable YAML. Handles nested objects, arrays, multiline strings, and special characters. No external dependencies.

  Paramètres, retour & exemple (2)

  Paramètres

  • input · string|object|array · requis
    The JSON to convert to YAML — a JSON string, or the value itself.
  • indent · number · optionnel
    Indentation size in spaces (default: 2)

  Exemple d'appel

  {"name": "json_to_yaml", "arguments": {"input": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"yaml": "texte", "lines": 1}}

• latency_benchmark≈ 342 tokens

  Measure response time of one or more HTTP endpoints (GET/POST). Runs N iterations and returns min/max/avg/p95 latency. Useful for API and MCP server benchmarking.

  Paramètres, retour & exemple (2)

  Paramètres

  • endpoints · string|array · requis
    Endpoints to benchmark. Accepts a single URL string, an array of URL strings, or an array of {url, method?, body?, headers?, label?} objects.
  • iterations · number · optionnel
    Number of iterations per endpoint (default: 3, max: 10)

  Exemple d'appel

  {"name": "latency_benchmark", "arguments": {"endpoints": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"results": [], "iterations": 1}}

• levenshtein_distance≈ 270 tokens

  Compute the Levenshtein (edit) distance and normalized similarity ratio between two strings. Supports batch comparison. Useful for fuzzy string matching, deduplication, and test result comparison.

  Paramètres, retour & exemple (4)

  Paramètres

  • a · string · optionnel
    First string (single-pair mode)
  • b · string · optionnel
    Second string (single-pair mode)
  • batch · array · optionnel
    Batch of {a,b} pairs (max 50)
  • case_insensitive · boolean · optionnel
    Ignore case differences (default: false)

  Exemple d'appel

  {"name": "levenshtein_distance", "arguments": {"a": "texte", "b": "texte", "batch": [], "case_insensitive": true}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"a": "texte", "b": "texte", "mode": "texte", "count": 1, "results": "texte", "distance": "texte", "similarity": "texte", "operations_needed": "texte"}}

• lint_commit_message≈ 234 tokens

  Validate a git commit message against the Conventional Commits spec (feat, fix, docs, style, refactor, test, chore, ci, perf, build). Returns compliance score, breaking change detection, and actionable suggestions.

  Paramètres, retour & exemple (2)

  Paramètres

  • strict · boolean · optionnel
    Enforce strict rules: max 72-char subject, imperative mood check (default: false)
  • message · string · requis
    Git commit message to validate

  Exemple d'appel

  {"name": "lint_commit_message", "arguments": {"message": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"type": "texte", "scope": "texte", "score": 1, "valid": true, "checks": "texte", "subject": "texte", "has_body": true, "is_breaking_change": true}}

• list_llm_models≈ 220 tokens

  List all LLM models available on ia-qa.com with their provider, API endpoint, and capabilities. Filter by provider name (e.g. "Groq", "HuggingFace", "OpenAI") or return the full catalog. Use this to discover which models are available before calling an LLM API, or to compare providers.

  Paramètres, retour & exemple (1)

  Paramètres

  • provider · string · optionnel
    Filter by provider name (case-insensitive). E.g. "Groq", "HuggingFace", "OpenAI", "Anthropic", "Google", "DeepSeek", "xAI", "Ollama". Omit for full catalog.

  Exemple d'appel

  {"name": "list_llm_models", "arguments": {"provider": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"total": 1, "filter": "texte", "models": "texte", "providers": "texte"}}

• list_local_tests≈ 172 tokens

  Discover .ia-eval.yaml LLM test suite files in the project directory. Scans CWD and standard sub-directories (evals/, tests/, contracts/). Returns file paths ready to pass to run_eval_contract.

  Paramètres, retour & exemple (1)

  Paramètres

  • dir · string · optionnel
    Directory to scan (defaults to server CWD)

  Exemple d'appel

  {"name": "list_local_tests", "arguments": {"dir": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"dir": "texte", "count": 1, "files": []}}

• llm_fit_finder≈ 417 tokens

  Find the best LLM for a given use case. Compares 30+ cloud API models and 12+ local models by cost, speed, benchmarks, features and VRAM requirements. Returns ranked recommendations with cost simulation. No API key needed.

  Paramètres, retour & exemple (8)

  Paramètres

  • mode · string · optionnel
    cloud (API models) or local (Ollama/self-hosted). Default: cloud
  • top_n · number · optionnel
    Number of recommendations to return (default: 5)
  • vram_gb · number · optionnel
    GPU VRAM in GB (only for mode=local). Default: 16
  • features · array · optionnel
    Required features: vision, function_calling, json_mode, streaming, reasoning
  • use_case · string · optionnel
    Primary use case: chatbot | code | rag | summarization | classification | reasoning | agents | multilingual
  • max_budget · number · optionnel
    Maximum monthly budget in USD (based on tokens_per_day)
  • quantization · string · optionnel
    Quantization (only for mode=local): Q4_K_M | Q8_0 | FP16. Default: Q4_K_M
  • tokens_per_day · number · optionnel
    Estimated daily token volume (default: 100000)

  Exemple d'appel

  {"name": "llm_fit_finder", "arguments": {"mode": "texte", "top_n": 1, "vram_gb": 1, "features": [], "use_case": "texte", "max_budget": 1}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"mode": "texte", "score": 1, "results": [], "vram_gb": "texte", "use_case": "texte", "quantization": "texte", "tokens_per_day": "texte", "total_matching": 1}}

• llm_format_check≈ 259 tokens

  Validate that an LLM output matches an expected format: JSON, Markdown, code block, bullet list, numbered list, table, YAML, XML, or custom regex. Essential for structured output testing.

  Paramètres, retour & exemple (3)

  Paramètres

  • output · string · requis
    The LLM output to validate
  • regex_pattern · string · optionnel
    Custom regex pattern (only when expected_format is "regex")
  • expected_format · string · requis · json | markdown_heading | code_block | bullet_list | numbered_list | table | yaml | xml
    Expected format

  Exemple d'appel

  {"name": "llm_format_check", "arguments": {"output": "texte", "expected_format": "json"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"valid": true, "checks": "texte", "failed": "texte", "passed": "texte", "total_checks": 1, "expected_format": "texte"}}

• llm_generate≈ 386 tokens

  Generate text using open-source LLM models hosted on Groq (ultra-fast) or HuggingFace Inference (serverless). No API key required — the server provides its own keys. Supported models: Qwen3 32B, Gemma 4 27B, Gemma 3 27B, Llama 3.3 70B, Llama 4 Scout, DeepSeek R1, Mistral Small 24B, and more. Use list_llm_models to see the full catalog. Rate-limited to prevent abuse.

  Paramètres, retour & exemple (5)

  Paramètres

  • model · string · optionnel
    Model ID (default: "openai/gpt-oss-20b"). Server-keyed whitelist only — Groq: openai/gpt-oss-20b, openai/gpt-oss-120b, qwen/qwen3.6-27b; HuggingFace: Qwen/Qwen3-32B, meta-llama/Llama-3.3-70B-Instruct, deepseek-ai/DeepSeek-R1, google/gemma-3-27b-it, and more. Other ids from list_llm_models are BYOK-only and will be rejected.
  • prompt · string · requis
    The user prompt / instruction to send to the model
  • system · string · optionnel
    Optional system prompt to set context or persona
  • max_tokens · number · optionnel
    Maximum tokens to generate (default: 2048, max: 4096)
  • temperature · number · optionnel
    Sampling temperature 0.0–1.5 (default: 0.7)

  Exemple d'appel

  {"name": "llm_generate", "arguments": {"prompt": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"model": "texte", "usage": "texte", "content": "texte", "provider": "texte", "latency_ms": 1}}

• llm_json_schema_check≈ 220 tokens

  Validate that an LLM JSON output matches a JSON Schema definition. Tests required fields, types, enums, nested objects, and arrays. Critical for function-calling and structured output testing.

  Paramètres, retour & exemple (2)

  Paramètres

  • output · string · requis
    The LLM JSON output (raw string, will be parsed)
  • schema · object · requis
    JSON Schema (draft-07 subset) to validate against

  Exemple d'appel

  {"name": "llm_json_schema_check", "arguments": {"output": "texte", "schema": {}}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"valid": true, "errors": [], "error_count": 1, "parse_error": "texte", "parsed_type": "texte"}}

• llm_output_validator≈ 432 tokens

  Validate an LLM response against QA criteria: format checks (JSON, code, markdown), content rules (must-include, must-not-include), length constraints, language detection, and safety patterns. Essential for QA testing LLM-powered features.

  Paramètres, retour & exemple (9)

  Paramètres

  • output · string · requis
    The LLM output text to validate
  • max_length · number · optionnel
    Maximum character length for the output
  • min_length · number · optionnel
    Minimum character length for the output
  • check_safety · boolean · optionnel
    Check for PII patterns (emails, phones, SSN), profanity signals, and prompt leakage
  • must_include · string · optionnel
    Comma-separated strings that MUST appear in the output
  • expected_format · string · optionnel · json | markdown | code | plain | any
    Expected output format
  • must_not_include · string · optionnel
    Comma-separated strings that must NOT appear (e.g. "TODO, FIXME, undefined, NaN")
  • check_json_schema · string · optionnel
    If expected_format is JSON, provide required keys as comma-separated list to validate the structure
  • expected_language · string · optionnel
    Expected language of the output (en, fr, es, de…). Checks for common words.

  Exemple d'appel

  {"name": "llm_output_validator", "arguments": {"output": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"total": 1, "checks": "texte", "failed": "texte", "passed": "texte", "verdict": "texte"}}

• lorem_ipsum≈ 218 tokens

  Generate Lorem Ipsum placeholder text for UI mockups, design prototypes, or test data population. Configurable paragraphs (1–10), sentences per paragraph (1–20), and approximate words per sentence (3–30).

  Paramètres, retour & exemple (3)

  Paramètres

  • paragraphs · number · optionnel
    Number of paragraphs to generate (1–10, default: 1)
  • words_per_sentence · number · optionnel
    Approximate words per sentence (3–30, default: 10)
  • sentences_per_paragraph · number · optionnel
    Sentences per paragraph (1–20, default: 5)

  Exemple d'appel

  {"name": "lorem_ipsum", "arguments": {"paragraphs": 1, "words_per_sentence": 1, "sentences_per_paragraph": 1}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"paragraphs": "texte", "paragraph_count": 1}}

• mcp_schema_lint≈ 200 tokens

  Lint an MCP tool definition for best practices: naming conventions, description quality, schema completeness, required fields consistency, description length. Returns actionable warnings.

  Paramètres, retour & exemple (1)

  Paramètres

  • tool_definition · object · requis
    MCP tool definition object with name, description, inputSchema

  Exemple d'appel

  {"name": "mcp_schema_lint", "arguments": {"tool_definition": {}}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"grade": "texte", "errors": "texte", "warnings": "texte", "error_count": 1, "quality_score": "texte", "warning_count": 1}}

• mcp_server_evaluate≈ 264 tokens

  Run a full compliance evaluation against a live MCP server URL. Tests: server reachability (ping), manifest discovery (GET /mcp), schema quality (snake_case names, descriptions, inputSchema), JSON-RPC 2.0 test call, and P50/P95 latency. Returns a PASS/FIX/BLOCK verdict with a 0-100 score and per-check details.

  Paramètres, retour & exemple (2)

  Paramètres

  • url · string · requis
    Base URL of the MCP server (e.g. https://www.ia-qa.com or http://localhost:3001)
  • test_tool_name · string · optionnel
    Specific tool name to use in the JSON-RPC test call (defaults to the first tool in the manifest)

  Exemple d'appel

  {"name": "mcp_server_evaluate", "arguments": {"url": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"url": "texte", "score": 1, "checks": {}, "latency": {}, "verdict": "texte"}}

• mcp_server_health_check≈ 250 tokens

  Generate a health check report for an MCP server's tool manifest. Validates tool definitions, schema quality, naming conventions, and documentation completeness. Paste the server manifest JSON to audit.

  Paramètres, retour & exemple (2)

  Paramètres

  • strict · boolean · optionnel
    Enable strict mode: also check for optional best practices (examples, default values, descriptions > 20 chars)
  • manifest · string · requis
    MCP server manifest JSON (the response from GET /mcp or tools/list)

  Exemple d'appel

  {"name": "mcp_server_health_check", "arguments": {"manifest": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"stats": {}, "total": 1, "checks": [], "failed": 1, "passed": 1, "verdict": "texte", "toolIssues": "texte"}}

• merge_json≈ 271 tokens

  Deep merge two JSON objects. Supports three array strategies: replace (default), concat, or unique (dedup concat). Nested objects are recursively merged — override takes precedence for primitives.

  Paramètres, retour & exemple (3)

  Paramètres

  • base · string|object|array · requis
    The base JSON object (merged into) — a JSON string, or the object itself.
  • override · string|object|array · requis
    The override JSON object (takes precedence) — a JSON string, or the object itself.
  • array_strategy · string · optionnel · replace | concat | unique
    Array merge strategy: replace (default), concat, or unique

  Exemple d'appel

  {"name": "merge_json", "arguments": {"base": [], "override": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"merged": "texte", "new_keys": [], "total_keys": 1, "overridden_keys": []}}

• metamorphic_check≈ 1222 tokens

  Reference-free stability primitive: instead of comparing an answer to a ground truth, it checks that an assistant's answer stays INVARIANT when the QUESTION is transformed (typo, casing, paraphrase, reordering, translation). Catches the failure class no reference answer can expose — an assistant that handles one phrasing well and a trivial variant of it badly. You bring the outputs (no model is called), so it is deterministic and free in tfidf mode. Relations: case (θ .95), typo (.90), paraphrase (.80), reorder (.80), translation (.75, embeddings only), specialization (.60, ADVISORY — directional, never gated). Returns PASS / FAIL / INVALID, where INVALID means the BASE answer was a refusal or too short so invariance was never measurable — an assistant that refuses every variant would otherwise score a perfect 1.0. Use run_semantic_tests alongside it: invariance without a correctness floor is a green light for a broken assistant.

  Paramètres, retour & exemple (7)

  Paramètres

  • base · object · requis
    The reference run: the original question and the answer your system produced for it.
  • mode · string · optionnel · tfidf | embeddings
    tfidf (default): free, lexical, deterministic — but a genuine paraphrase rarely reaches 0.80, so gate on case/typo and treat paraphrase as a trend. embeddings: OpenAI text-embedding-3-small, true semantic similarity, requires api_key. translation requires this mode.
  • api_key · string · optionnel
    OpenAI API key — required only when mode is embeddings.
  • variants · array · requis
    Answers produced for transformed versions of the same question, each tagged with the relation that was applied.
  • thresholds · object · optionnel
    Per-relation threshold overrides. Calibrate on your own corpus before gating — the defaults are starting points, not measurements.
  • require_all · boolean · optionnel
    If true (default), every gated variant must pass. KEEP THE DEFAULT for any run you gate on. Setting it false is not a tolerance dial but an off switch: relations have asymmetric pass rates (a typo variant usually scores ~1.0 because the answer really is identical), so one trivial row is enough to hold the whole run at PASS while a paraphrase fails. When that happens the result carries an explicit warning naming the failed rows.
  • baseline_guard · object · optionnel
    Correctness floor applied to the BASE answer before anything is scored. Failing it returns INVALID, not FAIL.

  Exemple d'appel

  {"name": "metamorphic_check", "arguments": {"base": {}, "variants": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"mode": "texte", "summary": {}, "verdict": "texte", "weakest": {}, "baseline": {}, "variants": [], "warnings": [], "thresholds": {}}}

• minify_js≈ 158 tokens

  Minify a JavaScript snippet, function, class, or module up to 50 KB using Terser. Returns minified code and byte savings. Use when embedding scripts in HTML templates, report payloads, or injecting inline code programmatically.

  Paramètres, retour & exemple (1)

  Paramètres

  • code · string · requis
    JavaScript code to minify (max 50kb)

  Exemple d'appel

  {"name": "minify_js", "arguments": {"code": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"minified": "texte"}}

• mock_from_schema≈ 249 tokens

  Generate realistic mock data from a JSON Schema. Supports all common types (string, number, integer, boolean, array, object, null), format hints (email, date, date-time, uri, uuid), enum, const, and nested schemas. Perfect for testing MCP tools with realistic data.

  Paramètres, retour & exemple (3)

  Paramètres

  • seed · string · optionnel
    Optional seed string for deterministic output (uses first char codes)
  • count · number · optionnel
    Number of mock objects to generate (default: 1, max: 20)
  • schema · string|object|array · requis
    The JSON Schema to generate from — a JSON string, or the schema object itself.

  Exemple d'appel

  {"name": "mock_from_schema", "arguments": {"schema": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"count": 1, "results": "texte"}}

• model_info≈ 183 tokens

  Get detailed specs for an AI model: context window, pricing per 1K tokens, knowledge cutoff, provider, multimodal support, reasoning capabilities, and feature list. Covers 30+ models from OpenAI, Anthropic, Google, DeepSeek, Meta, Mistral, Cohere, xAI.

  Paramètres, retour & exemple (1)

  Paramètres

  • model · string · requis
    Model name (e.g. "gpt-4o", "claude-3.5-sonnet", "gemini-2.5-pro")

  Exemple d'appel

  {"name": "model_info", "arguments": {"model": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"model": "texte", "pricing_per_1k": {}}}

• multimodal_eval_guide≈ 718 tokens

  Unified tool for multimodal AI evaluation: set action=guide for reference thresholds/interpretation (CLIP, FID, VQA), or set action=clip_score / fid_score / vqa_accuracy / pipeline to compute real metrics via HuggingFace Inference API and VLM BYOK calls. One tool for both reference and computation.

  Paramètres, retour & exemple (16)

  Paramètres

  • fid · object · optionnel
    [pipeline] {real_images, generated_images} for FID.
  • vqa · object · optionnel
    [pipeline] VQA config object (same inputs as vqa_accuracy).
  • clip · object · optionnel
    [pipeline] {image_url, text} for CLIP.
  • text · string · optionnel
    [clip_score only] Text description to compare against the image.
  • model · string · optionnel
    [vqa_accuracy] VLM model ID (default: gpt-4o).
  • score · number · optionnel
    [guide only] Optional score value to interpret.
  • action · string · optionnel · guide | clip_score | fid_score | vqa_accuracy | pipeline
    guide (default) = reference thresholds/interpretation. clip_score/fid_score/vqa_accuracy = compute that metric. pipeline = run all three.
  • metric · string · optionnel · clip_score | fid | vqa_accuracy | all
    [guide only] Metric to explain.
  • api_key · string · optionnel
    [vqa_accuracy] Your API key for the provider (BYOK).
  • image_url · string · optionnel
    [clip_score/vqa_accuracy] Public URL of the image.
  • test_cases · array · optionnel
    [vqa_accuracy] Array of {question, accepted_answers} objects.
  • real_images · array · optionnel
    [fid_score] Array of real image URLs.
  • image_base64 · string · optionnel
    [clip_score/vqa_accuracy] Base64-encoded image data.
  • system_prompt · string · optionnel
    [vqa_accuracy] Optional system prompt.
  • image_mime_type · string · optionnel
    [clip_score/vqa_accuracy] MIME type for base64 image.
  • generated_images · array · optionnel
    [fid_score] Array of generated image URLs.

  Exemple d'appel

  {"name": "multimodal_eval_guide", "arguments": {"fid": {}, "vqa": {}, "clip": {}, "text": "texte", "model": "texte", "score": 1}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"errors": "texte", "metrics": "texte", "results": "texte", "web_tool": "texte", "best_practices": [], "comparison_table": [], "score_interpretation": "texte"}}

• needle_haystack_generate≈ 323 tokens

  Generate a "needle in a haystack" test: embeds a target fact into a large block of filler text at a specified position. Use this to test LLM context window retrieval accuracy. Returns the full haystack, the question to ask, and metadata. No API key needed.

  Paramètres, retour & exemple (4)

  Paramètres

  • needle · string · requis
    The fact to hide (e.g. "The secret code is ALPHA-42")
  • tokens · integer · optionnel
    Target haystack size in tokens (default: 5000, max: 100000)
  • position · string · optionnel · start | middle | end | random
    Where to insert the needle: "start", "middle", "end", "random" (default: "middle")
  • question · string · requis
    The question to ask the LLM (e.g. "What is the secret code?")

  Exemple d'appel

  {"name": "needle_haystack_generate", "arguments": {"needle": "texte", "question": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"needle": "texte", "haystack": "texte", "position": "texte", "question": "texte", "insert_block": "texte", "total_blocks": 1, "estimated_tokens": "texte"}}

• normalize_vector≈ 242 tokens

  L2-normalize a float vector (produce a unit vector with norm=1). Required by many vector DBs (Pinecone, Qdrant cosine). Supports batch normalization of up to 1000 vectors.

  Paramètres, retour & exemple (2)

  Paramètres

  • batch · array · optionnel
    Batch of vectors to normalize (overrides vector)
  • vector · array · optionnel
    Single vector to normalize

  Exemple d'appel

  {"name": "normalize_vector", "arguments": {"batch": [], "vector": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"mode": "texte", "norm": 1, "count": 1, "index": "texte", "vector": "texte", "results": "texte", "dimension": 1, "norm_after": 1, "normalized": "texte", "norm_before": 1}}

• normalize_whitespace≈ 331 tokens

  Normalize whitespace: trim trailing spaces, collapse blank lines, normalize line endings (LF/CRLF), convert tabs to spaces. Useful for cleaning code, configs, and text before processing.

  Paramètres, retour & exemple (7)

  Paramètres

  • input · string · requis
    Text to normalize
  • trim_file · boolean · optionnel
    Trim leading/trailing blank lines (default: true)
  • trim_lines · boolean · optionnel
    Trim trailing whitespace from each line (default: true)
  • line_ending · string · optionnel
    "lf" (default), "crlf", or "cr"
  • tab_to_spaces · number · optionnel
    Convert tabs to N spaces (omit to keep tabs)
  • collapse_blanks · boolean · optionnel
    Collapse runs of blank lines down to max_blank_lines (default: true)
  • max_blank_lines · number · optionnel
    Blank lines to keep when collapsing, 0-10 (default: 2)

  Exemple d'appel

  {"name": "normalize_whitespace", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"result": "texte", "line_ending": "texte", "original_length": 1, "normalized_length": 1}}

• number_base_convert≈ 230 tokens

  Convert numbers between bases: decimal, binary, octal, hexadecimal, or any base 2–36. Auto-detects 0x, 0b, 0o prefixes.

  Paramètres, retour & exemple (3)

  Paramètres

  • input · string · requis
    Number to convert (e.g., "255", "0xFF", "0b1010", "0o77")
  • to_base · number · optionnel
    Target base 2–36 (omit to get all common bases)
  • from_base · number · optionnel
    Source base 2–36 (auto-detects prefix if omitted)

  Exemple d'appel

  {"name": "number_base_convert", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"octal": "texte", "binary": "texte", "result": "texte", "decimal": "texte", "to_base": "texte", "from_base": "texte", "hexadecimal": "texte"}}

• openapi_validate≈ 257 tokens

  Validate the structure of an OpenAPI 3.x specification (JSON or YAML). Checks required top-level fields (openapi, info.title, info.version, paths), validates each operation (responses, operationId uniqueness), detects undeclared $ref components, and flags missing 2xx responses. Returns a PASS/FAIL verdict, a 0–100 compliance score, and a list of errors and warnings with JSON-pointer locations. Use before publishing an API spec or generating SDK code.

  Paramètres, retour & exemple (1)

  Paramètres

  • input · string|object|array · requis
    The OpenAPI 3.x spec — a JSON string, a YAML string, or the already-parsed spec object.

  Exemple d'appel

  {"name": "openapi_validate", "arguments": {"input": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"score": 1, "stats": {}, "errors": "texte", "verdict": "texte", "warnings": "texte"}}

• optimize_prompt_tokens≈ 262 tokens

  Compress an LLM prompt by removing filler words, verbose phrases, duplicate sentences, and unnecessary whitespace. Returns optimized text with token savings breakdown. 100% deterministic, no API key needed.

  Paramètres, retour & exemple (2)

  Paramètres

  • text · string · requis
    The prompt text to optimize
  • options · object · optionnel
    Toggle optimization steps (all true by default)

  Exemple d'appel

  {"name": "optimize_prompt_tokens", "arguments": {"text": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"steps": "texte", "optimized": "texte", "tokens_after": "texte", "tokens_saved": "texte", "percent_saved": "texte", "tokens_before": "texte"}}

• parse_csv≈ 284 tokens

  Parse a CSV string into a JSON array of objects (or raw arrays). Full RFC 4180: quoted fields may contain the delimiter, embedded newlines (the Excel/Sheets multi-line cell), and doubled quotes. Custom delimiters supported. An unterminated quote is rejected with its position rather than parsed into corrupted rows. Use when processing spreadsheet exports, data imports, or structured text pipelines where the source is CSV. Supports up to 200 KB.

  Paramètres, retour & exemple (3)

  Paramètres

  • input · string · requis
    CSV content to parse
  • header · boolean · optionnel
    Treat the first row as headers (default: true)
  • delimiter · string · optionnel
    Field delimiter character (default: ",")

  Exemple d'appel

  {"name": "parse_csv", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"rows": [], "columns": 1, "headers": [], "row_count": 1}}

• parse_http_headers≈ 208 tokens

  Parse a raw HTTP headers block into a structured JSON object. Detects multi-value headers, masks Authorization values, and optionally audits for missing security headers (HSTS, CSP, X-Frame-Options, etc.).

  Paramètres, retour & exemple (2)

  Paramètres

  • headers · string · requis
    Raw HTTP headers (one "Name: Value" per line)
  • analyze_security · boolean · optionnel
    Audit for missing security headers (default: true)

  Exemple d'appel

  {"name": "parse_http_headers", "arguments": {"headers": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"parsed": {}, "security": {}, "header_count": 1}}

• post_jira_comment≈ 287 tokens

  Post the output of jira_to_test_suite as a formatted comment on the source Jira ticket. Converts Gherkin, E2E steps, API tests, and ambiguities into Atlassian Document Format (ADF). STATEFUL — creates a comment on the issue.

  Paramètres, retour & exemple (5)

  Paramètres

  • issue_key · string · requis
    Jira issue key, e.g. "PROJ-123"
  • jira_email · string · requis
    Atlassian account email
  • jira_token · string · requis
    Atlassian API token
  • test_suite · object · requis
    The test_suite object from jira_to_test_suite result
  • jira_base_url · string · requis
    Atlassian base URL

  Exemple d'appel

  {"name": "post_jira_comment", "arguments": {"issue_key": "texte", "jira_email": "texte", "jira_token": "texte", "test_suite": {}, "jira_base_url": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"success": true, "comment_id": "texte", "comment_url": "texte"}}

• pr_gatekeeper≈ 308 tokens

  Compound quality gate for pull requests. Runs three sequential checks: (1) secret detection — scans diff for API keys, tokens, passwords matching 16 regex patterns; (2) bug analysis — heuristic scan for eval(), innerHTML, empty catch, console.log, TODO/FIXME; (3) commit message linting against Conventional Commits spec. Returns gate verdict (PASS/WARN/BLOCK), blockers, and actionable warnings. Use before merging any code change.

  Paramètres, retour & exemple (3)

  Paramètres

  • diff · string · requis
    Unified git diff (output of `git diff HEAD`)
  • context · string · optionnel
    Optional: PR title or description for richer bug analysis
  • commit_message · string · requis
    The commit message to lint (e.g. "feat(auth): add OAuth2 login")

  Exemple d'appel

  {"name": "pr_gatekeeper", "arguments": {"diff": "texte", "commit_message": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"flags": [], "score": 1, "checks": {}, "verdict": "texte"}}

• prompt_injection_scan≈ 399 tokens

  Scan user input or prompts for common prompt injection patterns. Detects system prompt overrides, jailbreak attempts, role manipulation, encoding tricks, delimiter attacks (chat-template tags `<|im_start|>`/`[INST]`/`<<SYS>>` AND fake role headers imitating markdown or chat separators: "### System:", "--- SYSTEM ---", "---BEGIN SYSTEM OVERRIDE---", "--- SYSTEM:"), template/interpolation injection ({{...}}, ${...}), and context-exfiltration attempts ("repeat everything above"). A match inside quoted or fenced text (documentation citing a payload) is reported one severity level lower and marked `quoted` — never suppressed, since an LLM reading the document as data can still follow a quoted instruction. A quote preceded by a live imperative ("output the following: ...") keeps its full severity.

  Paramètres, retour & exemple (2)

  Paramètres

  • input · string · requis
    The user input or prompt to scan for injection patterns
  • sensitivity · string · optionnel · low | medium | high
    Detection sensitivity (default: medium)

  Exemple d'appel

  {"name": "prompt_injection_scan", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"notes": "texte", "detections": "texte", "risk_level": "texte", "sensitivity": "texte", "input_length": 1, "detections_count": 1, "quoted_detections": 1, "injection_detected": true}}

• prompt_template_fill≈ 234 tokens

  Fill a prompt template with variables. Supports {{variable}} syntax and {{#if key}}...{{/if}} conditional blocks. Returns the filled prompt and lists unfilled variables.

  Paramètres, retour & exemple (3)

  Paramètres

  • strict · boolean · optionnel
    Throw error if any variable is not provided (default: false)
  • template · string · requis
    Prompt template with {{variable}} placeholders
  • variables · object · optionnel
    Key-value pairs to fill (e.g. {"name":"Alice","role":"engineer"})

  Exemple d'appel

  {"name": "prompt_template_fill", "arguments": {"template": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"result": "texte", "total_vars": "texte", "filled_variables": "texte", "unfilled_variables": "texte"}}

• prompt_test_suite≈ 418 tokens

  Define a test suite for a prompt: provide the system prompt, user prompt, and expected output criteria. Returns a test plan with scored rubric — use this as input for manual or automated LLM evaluation.

  Paramètres, retour & exemple (10)

  Paramètres

  • max_tokens · number · optionnel
    Max token budget for the test
  • temperature · number · optionnel
    Temperature to use
  • user_prompt · string · requis
    The user prompt to send
  • check_safety · boolean · optionnel
    Include safety/PII checks in the rubric
  • must_include · string · optionnel
    Required content (comma-separated)
  • system_prompt · string · requis
    The system prompt under test
  • expected_format · string · optionnel · json | markdown | code | plain | any
    Expected output format
  • must_not_include · string · optionnel
    Forbidden content (comma-separated)
  • expected_behavior · string · optionnel
    Description of what the LLM should do (free text)
  • adversarial_prompts · boolean · optionnel
    Auto-generate adversarial test variants (jailbreak, injection, edge cases)

  Exemple d'appel

  {"name": "prompt_test_suite", "arguments": {"user_prompt": "texte", "system_prompt": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"rubric": "texte", "categories": [], "total_tests": 1, "instructions": "texte", "test_suite_name": "texte"}}

• rag_relevance_rank≈ 236 tokens

  Rank an array of text chunks by relevance to a query using TF-IDF scoring. Simulates retrieval ranking for RAG testing without needing embeddings or an API.

  Paramètres, retour & exemple (3)

  Paramètres

  • query · string · requis
    The user query
  • top_k · number · optionnel
    Return top K results (default: all)
  • chunks · array · requis
    Array of text chunks to rank

  Exemple d'appel

  {"name": "rag_relevance_rank", "arguments": {"query": "texte", "chunks": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"rank": 1, "index": "texte", "query": "texte", "score": "texte", "results": "texte", "returned": 1, "total_chunks": 1, "chunk_preview": "texte", "keyword_overlap": "texte"}}

• rate_tool≈ 416 tokens

  Give honest usage feedback on an IA-QA MCP tool. Provide a score (1-5) and a comment. Rate low (1-2) if the tool was wrong, irrelevant, or a poor fit; rate high (4-5) only if it genuinely solved your need. Ratings are aggregated on a public dashboard at /devtools/mcp-ratings. Skip rating routine successes — we want signal, not praise. Example: rate_tool({ tool_name: "format_json", score: 2, comment: "Tried to pretty-print a JSON5 file, it rejected trailing commas — not usable for my case." })

  Paramètres, retour & exemple (3)

  Paramètres

  • score · number · requis
    Rating from 1 (poor) to 5 (excellent)
  • comment · string · optionnel
    Strongly encouraged — explain what you were trying to do and whether the tool got you there. Be specific about what was missing, wrong, or a poor fit. This is the most valuable part of the rating. Up to 2000 chars are stored; go over and the response says so (truncated: true) — send the remainder as a second call rather than assuming it landed.
  • tool_name · string · requis
    Name of the MCP tool to rate (e.g. "format_json", "shield_analyze")

  Exemple d'appel

  {"name": "rate_tool", "arguments": {"score": 1, "tool_name": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"ok": true, "score": "texte", "comment": "texte", "message": "texte", "rated_at": "texte", "tool_name": "texte", "truncated": true, "stored_chars": 1}}

• redact_pii≈ 254 tokens

  Automatically detect and redact Personally Identifiable Information (PII) from text. Replaces emails, phone numbers, SSNs, credit cards, IP addresses, and JWT tokens with [REDACTED_TYPE] placeholders. Safe to use before logging or sending to an LLM.

  Paramètres, retour & exemple (3)

  Paramètres

  • input · string · requis
    Text to redact PII from
  • types · string · optionnel
    Comma-separated types to redact (default: all). Options: email, phone, ssn, credit_card, ip_address, jwt
  • marker · string · optionnel
    Custom replacement marker (default: "REDACTED"). Result: [REDACTED_EMAIL]

  Exemple d'appel

  {"name": "redact_pii", "arguments": {"input": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"clean": true, "pii_found": "texte", "replacements": "texte", "redacted_text": "texte", "total_redactions": "texte"}}

• regex_test≈ 330 tokens

  Test a regular expression pattern against an input string and return all matches with their index positions and named capture groups. Use for validating user inputs, extracting structured data from text, or debugging regex patterns. Supports flags g, i, m, s, u, y. The match runs in an isolated thread with a 500 ms budget: a pattern that blows up (catastrophic backtracking, e.g. "(a+)+$") comes back as redos_detected:true — a real ReDoS verdict on your pattern — instead of hanging.

  Paramètres, retour & exemple (3)

  Paramètres

  • flags · string · optionnel
    Regex flags: g (global), i (case-insensitive), m (multiline), s (dotAll) — default: ""
  • input · string · requis
    The string to test against (max 50 KB)
  • pattern · string · requis
    Regular expression pattern (without delimiters)

  Exemple d'appel

  {"name": "regex_test", "arguments": {"input": "texte", "pattern": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"note": "texte", "flags": "texte", "matched": "texte", "matches": "texte", "pattern": "texte", "verdict": "texte", "elapsed_ms": 1, "match_count": 1, "redos_detected": true}}

• rerank_evaluate≈ 651 tokens

  Evaluate RAG retrieval quality: rank passages against a query and compute Precision@k / Recall@k plus a PASS/FAIL CI verdict from ground-truth relevance labels. Three modes, all keyless except the last. (1) BYO scores — give each passage the `score` your own reranker produced (Cohere, Jina, a self-hosted NIM, a cross-encoder): deterministic, offline, and it evaluates YOUR reranker rather than someone else's. This is the mode to gate CI on. (2) Default, no scores and no key — ranks with local BM25, a lexical keyword baseline: it answers "does a keyword floor already surface my relevant passages?", never "is my neural reranker good". (3) Live NVIDIA reranker — supply `api_key` for an NVIDIA account that still has reranking entitlement; NVIDIA retired its hosted reranking endpoints on 2026-05-18, so this one is for accounts that were grandfathered in.

  Paramètres, retour & exemple (5)

  Paramètres

  • query · string · requis
    The search query or question to rank against
  • top_k · integer · optionnel
    k for Precision@k evaluation (default 3)
  • api_key · string · optionnel
    Your NVIDIA API key (BYOK), used only when no passage carries a score. Transits RAM for the single call, never stored.
  • passages · array · requis
    Array of passage objects to rank (min 2, max 20)
  • threshold · number · optionnel
    Minimum Precision@k to PASS (0-1, default 0.5)

  Exemple d'appel

  {"name": "rerank_evaluate", "arguments": {"query": "texte", "passages": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"k": 1, "mode": "texte", "model": "texte", "ranked": [], "verdict": "texte", "threshold": 1, "latency_ms": 1, "recall_at_k": 1, "precision_at_k": 1, "total_passages": 1, "has_ground_truth": true}}

• response_quality_score≈ 379 tokens

  Score an LLM response against the criteria you pass: coverage of expected_keywords and compliance with max_length. Returns a 0-100 score over the criteria actually measured — or total_score null with the reason when none is given, because relevance and correctness cannot be read off word overlap (a correct one-word answer shares no words with its question). For meaning against a reference answer, use run_semantic_tests. Also returns unscored signals: question-term overlap, average sentence length, markdown.

  Paramètres, retour & exemple (4)

  Paramètres

  • question · string · requis
    The original question/prompt
  • response · string · requis
    The LLM response to score
  • max_length · number · optionnel
    Maximum character length. Scored as a proportional penalty beyond it.
  • expected_keywords · array · optionnel
    Keywords a correct answer must contain (case-insensitive). Scored as coverage.

  Exemple d'appel

  {"name": "response_quality_score", "arguments": {"question": "texte", "response": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"grade": "texte", "stats": {}, "reason": "texte", "signals": {}, "breakdown": {}, "max_score": 1, "total_score": "texte", "not_measured": []}}

• run_eval_contract≈ 431 tokens

  Parse a .ia-eval.yaml LLM test suite, call the specified LLM model for each scenario, run all configured scorers, and return a structured JSON report with per-scenario Pass/Fail verdicts and a Markdown summary. Use list_local_tests to discover available test files.

  Paramètres, retour & exemple (4)

  Paramètres

  • api_keys · object · optionnel
    API keys to use for LLM generation (all optional — falls back to server env vars)
  • overrides · object · optionnel
    Override contract defaults
  • contract_path · string · optionnel
    Absolute or relative path to a .ia-eval.yaml file (required unless inline_contract is provided)
  • inline_contract · object · optionnel
    Raw contract object (alternative to contract_path). Must contain top-level "metadata" ({name, version, model?, provider?}), "expectations" ({min_score?}), and "scenarios" ([{id, input, ground_truth?}]) — scenarios alone are rejected. Use generate_eval_yaml to scaffold one.

  Exemple d'appel

  {"name": "run_eval_contract", "arguments": {"api_keys": {}, "overrides": {}, "contract_path": "texte", "inline_contract": {}}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"summary": {}, "metadata": {}, "warnings": "texte", "contract_path": "texte", "scenario_results": "texte"}}

• run_pr_gate_pipeline≈ 392 tokens

  Review triage for a pull request. Takes a unified git diff (`git diff HEAD`) and returns: diff-lint findings with the lines that produced them, regression impact areas, a risk score 0–100 with the factors that built it (churn, files touched, sensitive paths, whether any test file changed, lint severities, impacted risk areas), generated test cases, and a PASS / CONDITIONAL / BLOCK recommendation. Advisory: the score measures properties of the diff, not the correctness of the change — it does not read the code semantically and does not replace a reviewer or a static analyser. See notAnalysed in the response.

  Paramètres, retour & exemple (2)

  Paramètres

  • context · string · optionnel
    Optional PR title or description for richer analysis
  • git_diff · string · requis
    Unified git diff (output of `git diff HEAD` or copied from GitHub diff view)

  Exemple d'appel

  {"name": "run_pr_gate_pipeline", "arguments": {"git_diff": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"sla": "texte", "high": "texte", "topBugs": [], "critical": "texte", "bugsFound": 1, "riskLevel": "texte", "riskScore": "texte", "disclaimer": "texte", "impactAreas": [], "inputFormat": "texte", "notAnalysed": [], "riskFactors": []}}

• run_semantic_tests≈ 537 tokens

  Semantic assertion primitive: compare actual vs expected text pairs using cosine similarity + ROUGE-L. Two modes: tfidf (default, free, no API key) or embeddings (OpenAI text-embedding-3-small, BYOK, true semantic similarity). Returns per-case PASS/FAIL verdicts and an overall verdict. CI-ready: pipe the JSON verdict field to gate a build.

  Paramètres, retour & exemple (5)

  Paramètres

  • mode · string · optionnel · tfidf | embeddings
    tfidf (default): fast, free, lexical. embeddings: OpenAI text-embedding-3-small, true semantic similarity, requires api_key.
  • cases · array · requis
    Array of (actual, expected) pairs to evaluate.
  • api_key · string · optionnel
    OpenAI API key — required only when mode is embeddings.
  • thresholds · object · optionnel
    Pass/fail thresholds (defaults: cosine 0.75, rouge_l 0.5).
  • require_all · boolean · optionnel
    If true (default), all cases must pass for overall PASS. If false, at least one case passing returns PASS.

  Exemple d'appel

  {"name": "run_semantic_tests", "arguments": {"cases": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"mode": "texte", "total": 1, "failed": 1, "passed": 1, "results": [], "verdict": "texte"}}

• run_vlm_test_suite≈ 666 tokens

  Run a test suite against a Vision-Language Model (VLM) — send an image (URL or base64) + N test cases (each with a question + assertion) to GPT-4o, Claude 3.5, or Gemini. Returns per-case PASS/FAIL verdicts, a pass rate, an overall PASS/WARNING/FAIL verdict (customizable threshold), and latency stats. Assertion types: contains, not_contains, json_format, min_length, max_length, semantic_contains (TF-IDF cosine similarity ≥ 0.4). BYOK: requires your own API key for the target provider.

  Paramètres, retour & exemple (8)

  Paramètres

  • model · string · requis · gpt-4o | gpt-4o-mini | claude-3-5-sonnet-20241022 | claude-3-5-haiku-20241022 | gemini-1.5-flash | gemini-2.0-flash
    VLM model to use.
  • api_key · string · requis
    API key for the model provider (OpenAI sk-, Anthropic sk-ant-, or Google AIzaSy...).
  • image_url · string · optionnel
    Public URL of the image to evaluate (required unless image_base64 is provided).
  • threshold · number · optionnel
    Pass rate threshold for overall verdict (default: 80, 0–100).
  • test_cases · array · requis
    Array of test cases to run.
  • image_base64 · string · optionnel
    Base64-encoded image data (required unless image_url is provided).
  • system_prompt · string · optionnel
    Optional system prompt sent to the VLM.
  • image_mime_type · string · optionnel
    MIME type of the image if using image_base64 (default: image/jpeg).

  Exemple d'appel

  {"name": "run_vlm_test_suite", "arguments": {"model": "gpt-4o", "api_key": "texte", "test_cases": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"model": "texte", "total": 1, "failed": 1, "passed": 1, "results": [], "verdict": "texte"}}

• run_vlm_test_suite_batch≈ 657 tokens

  Compare multiple VLMs on the same test suite in parallel — send an image (URL or base64) + N test cases to all models simultaneously. Returns per-model PASS/FAIL verdicts, pass rates, latency stats, and a comparison table. Assertion types: contains, not_contains, json_format, min_length, max_length, semantic_contains. BYOK: requires API keys for each provider.

  Paramètres, retour & exemple (8)

  Paramètres

  • models · array · requis
    Array of model IDs to compare (runs in parallel).
  • api_keys · object · requis
    Map of model ID → API key. Example: { "gpt-4o": "sk-...", "claude-3-5-sonnet-20241022": "sk-ant-..." }
  • image_url · string · optionnel
    Public URL of the image to evaluate (required unless image_base64 is provided).
  • threshold · number · optionnel
    Pass rate threshold for overall verdict (default: 80, 0–100).
  • test_cases · array · requis
    Array of test cases to run against every model.
  • image_base64 · string · optionnel
    Base64-encoded image data (required unless image_url is provided).
  • system_prompt · string · optionnel
    Optional system prompt sent to every VLM.
  • image_mime_type · string · optionnel
    MIME type of the image if using image_base64 (default: image/jpeg).

  Exemple d'appel

  {"name": "run_vlm_test_suite_batch", "arguments": {"models": [], "api_keys": {}, "test_cases": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"suites": [], "verdict": "texte", "total_failed": 1, "total_passed": 1}}

• sandbox_scenario≈ 492 tokens

  Get a ready-made selector-drift test case with a known-correct answer, for testing this MCP server or an agent workflow end to end. Each scenario is a real DOM capture of a deliberately breakable app, taken before and after one specific UI change (a renamed label, two swapped buttons, a duplicated locator, an element moved behind a menu…), plus the verdict those two contracts MUST produce. Call with no arguments to list the scenarios; call with a scenario id to get "baseline" and "current" mappings. THE LOOP: pass baseline and current to diff_mappings, then compare its "verdict" and "counts" to this tool's "expected" — they must match exactly. A mismatch means this server's diff engine has drifted, not that your inputs are wrong. Deterministic and offline: the captures are committed fixtures, identical on every call. Try it live at https://www.ia-qa.com/devtools/sandbox

  Paramètres, retour & exemple (2)

  Paramètres

  • scenario · string · optionnel
    Scenario id. Omit to list every available scenario with its expected verdict. Ids: no-change, swap-label, add-testid, duplicate-role-name, remove-element, insert-sibling, rename-label, counter-label, move-behind-menu, add-element
  • include_html · boolean · optionnel
    Include the generated HTML of the mutated page (default false). Only useful if you want to render or re-capture it yourself; the loop does not need it.

  Exemple d'appel

  {"name": "sandbox_scenario", "arguments": {"scenario": "texte", "include_html": true}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"html": "texte", "page": "texte", "blurb": "texte", "count": 1, "title": "texte", "current": {}, "teaches": "texte", "baseline": {}, "expected": {}, "scenario": "texte", "mutations": [], "scenarios": []}}

• score_geo_signals≈ 205 tokens

  Analyze a webpage <head> HTML (or full HTML) for GEO (Generative Engine Optimization) signals. Returns a score /60 with per-check results and improvement tips. GEO = optimizing pages for AI-powered search engines (ChatGPT Search, Perplexity, etc.).

  Paramètres, retour & exemple (1)

  Paramètres

  • head_html · string · requis
    Raw HTML of the <head> section (or full page HTML) to analyze

  Exemple d'appel

  {"name": "score_geo_signals", "arguments": {"head_html": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"grade": "texte", "score": "texte", "checks": "texte", "passed": 1, "max_score": 1, "total_checks": 1}}

• search_jira_issues≈ 357 tokens

  Search Jira using JQL (Jira Query Language). Returns matching issues with key fields. Ideal for finding open bugs, sprint tickets, or issues by label/assignee/component. BYOK — credentials transit in-memory only, never stored.

  Paramètres, retour & exemple (6)

  Paramètres

  • jql · string · requis
    JQL query string, e.g. "project = PROJ AND status = Open AND assignee = currentUser() ORDER BY priority DESC"
  • fields · array · optionnel
    Fields per issue. Default: summary, status, assignee, priority, issuetype, labels, created, updated
  • jira_email · string · requis
    Atlassian account email
  • jira_token · string · requis
    Atlassian API token
  • max_results · number · optionnel
    Max issues to return (default: 10, max: 50)
  • jira_base_url · string · requis
    Atlassian base URL, e.g. "https://mycompany.atlassian.net"

  Exemple d'appel

  {"name": "search_jira_issues", "arguments": {"jql": "texte", "jira_email": "texte", "jira_token": "texte", "jira_base_url": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}], "structuredContent": {"jql": "texte", "total": 1, "issues": [], "returned": 1}}

+ 32 autres outils (affichage limité à 120).

C'est votre serveur ?

Revendiquez-le sur l'espace créateur (preuve GitHub, domaine ou npm) pour afficher le badge « propriétaire vérifié » et lancer un pentest comportemental à la demande afin d'obtenir le badge « Vérifié ».

Créer un compte / revendiquer →

Source

https://www.ia-qa.com/mcp