IA-QA — 130+ QA & Dev Tools for AI Agents

metamorphic_check

Read-onlyIdempotent

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.

Input Schema

TableJSON Schema

Name	Required	Description
`base`	Yes	The reference run: the original question and the answer your system produced for it.
`mode`	No	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`	No	OpenAI API key — required only when mode is embeddings.
`variants`	Yes	Answers produced for transformed versions of the same question, each tagged with the relation that was applied.
`thresholds`	No	Per-relation threshold overrides. Calibrate on your own corpus before gating — the defaults are starting points, not measurements.
`require_all`	No	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`	No	Correctness floor applied to the BASE answer before anything is scored. Failing it returns INVALID, not FAIL.

Output Schema

TableJSON Schema

Name	Required	Description	Default
`mode`	No
`summary`	No
`verdict`	No
`weakest`	No
`baseline`	No
`variants`	No
`warnings`	No
`thresholds`	No

TDQS

A4.9/5.0

Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already declare readOnlyHint=true, idempotentHint=true, and destructiveHint=false, and the description adds substantial behavioral context beyond that: deterministic and free in tfidf mode, no model call, PASS/FAIL/INVALID semantics, INVALID meaning base refusal/too short, relation-specific thresholds, and the require_all off-switch warning. No contradiction with annotations.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is dense and long, but every sentence carries useful information for a complex 7-parameter tool. It is front-loaded with the core concept and then covers thresholds, return values, and guardrails. Slightly less structured than ideal, but not bloated or redundant.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's complexity (nested objects, enums, output schema, 7 params), the description is remarkably complete: it explains return values, mode requirements, relation thresholds, the require_all warning, baseline_guard behavior, and the companion tool. It leaves no major behavioral gap for an agent to discover at runtime.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, but the description adds meaning well beyond the schema: it explains mode differences (tfidf vs embeddings), relation thresholds and which are advisory, the require_all 'off switch' behavior with asymmetric pass rates, and the baseline_guard first-200-chars refusal matching. This materially helps an agent choose and set parameters correctly.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description opens with 'Reference-free stability primitive' and immediately states the specific behavior: 'checks that an assistant's answer stays INVARIANT when the QUESTION is transformed.' This clearly distinguishes it from ground-truth comparison tools and from sibling tools like run_semantic_tests. The verb+resource+scope is precise and memorable.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description explicitly frames when to use it ('instead of comparing an answer to a ground truth'), what the caller must bring ('You bring the outputs (no model is called)'), and names the companion tool: 'Use run_semantic_tests alongside it: invariance without a correctness floor is a green light for a broken assistant.' This gives clear context and an explicit alternative/complement.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All IA-QA — 130+ QA & Dev Tools for AI Agents tools Inspect

Try in Browser

RayrunBuild, deploy, and operate MCP servers without becoming an infrastructure team.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

TDQS

A3.6/5.0

Disambiguation2/5

Multiple tools overlap significantly: compare_models/llm_fit_finder/model_info/list_llm_models all compare models; similarity_score/embedding_similarity/run_semantic_tests all measure text similarity; detect_secrets/secret_scan/analyze_diff_bugs/pr_gatekeeper all scan for secrets. Descriptions attempt to differentiate, but the boundaries between many tools are unclear, making selection error-prone.

Naming Consistency4/5

The vast majority of tools follow a snake_case verb_noun pattern (validate_email, generate_uuid, parse_csv), making the set mostly predictable. A few notable deviations exist (pr_gatekeeper, llm_fit_finder, cot_analyzer, jira_to_test_suite, needle_haystack_generate) but they are the exception rather than the rule.

Tool Count1/5

With 149 tools, this set is far beyond the 50+ threshold for an extreme mismatch. Even as a general-purpose QA & Dev toolkit, the sheer number overwhelms and exceeds any reasonable scope, making discovery and selection impractical.

Completeness4/5

The toolkit covers an impressively broad range: text processing, LLM evaluation, security auditing, web checks, MCP validation, Jira/Confluence integration, and more. Minor gaps exist, such as missing delete/update for webhooks and Confluence pages, and no create/update for Jira issues, but these are workable around.

Resources

Need Help?