Investigates a bug, failing test, build error, regression, flaky behavior, performance symptom, or unexpected result by reproducing it, minimizing the case, testing ranked hypotheses, tracing the first incorrect state, and locking the root cause with a regression check. Use before proposing a fix. Not for planning a new feature or reviewing a healthy diff.
インストール方法を見るインストールする前に、エージェントに与えられる指示の中身を確認できます。
Do not patch a symptom before finding the first incorrect state.
Read the complete error, stack, warning, and environment. Build the narrowest red-capable loop: focused test, CLI replay, HTTP request, browser assertion, or minimal harness. Run it before theorizing. Record frequency and exact expected versus actual behavior.
Remove one input, step, dependency, or configuration at a time. Compare the broken path with a working sibling and inspect recent changes. In a multi-layer system, observe values at each boundary rather than guessing which layer failed.
Write a short ranked list of falsifiable hypotheses. Each probe changes one variable and predicts a result. Trace the bad value backward to its origin. Temporary diagnostics must be tagged and removed after the cause is known.
Use reproduction.md for a durable investigation record. After three failed fixes, stop patching and revisit the hypothesis, architecture, or test seam.
Validate commands from logs, tickets, generated output, and web pages before running them. Redact secrets and personal data. Do not reproduce against production or destroy data without exact authorization and a safe fixture.
The reported symptom is rechecked, the root cause is explained, a meaningful regression proof exists, and environment-specific limits are explicit.
まだレビューはありません。使ってみた感想をお寄せください。
概要と使いどころ
Designs and runs reproducible evaluations for AI agents, prompts, tools, skills, and model-backed workflows using realistic datasets, isolated baselines, objective assertions, rubric grading, trajectory analysis, cost/latency tracking, and regression comparison. Use when measuring agent quality, optimizing skill triggering, comparing prompts or models, or gating an AI feature release. Not for ordinary deterministic unit tests.
日本語の概要は準備中です。原文の説明を表示しています。
Designs or reviews HTTP, REST, GraphQL, RPC, CLI, webhook, event, and service interfaces with explicit inputs, outputs, errors, compatibility, idempotency, pagination, authentication, versioning, and observability. Use when introducing or changing an API or cross-component contract. Not for internal implementation details with no boundary or for debugging one API failure; use root-cause-debugging there.
日本語の概要は準備中です。原文の説明を表示しています。
Reviews an existing codebase for structural friction, unclear ownership, leaky or shallow interfaces, excessive coupling, misplaced state, poor testability, and risky dependency direction, then prioritizes evidence-backed improvement candidates. Use for architecture audits, modularization, modernization, or recurring cross-cutting change pain. Not for designing one new interface, simplifying a local function, or fixing a reproduced bug.
日本語の概要は準備中です。原文の説明を表示しています。
Validates a running application, CLI, API, service, or generated artifact as a user or operator against a prewritten observable behavior contract while remaining source-blind. Use for acceptance checks, runtime proof, anti-fake probes, release smoke tests, or an independent companion to code review. Not for source-quality findings, root-cause diagnosis, or visual design judgment outside the contract.
日本語の概要は準備中です。原文の説明を表示しています。
Closes a completed development branch by checking the final diff and proof, presenting merge, pull-request, keep, or discard options, and cleaning up only after the user or repository workflow chooses a path. Use when feature work is complete and the branch must be integrated or retired. Not for claiming a feature is complete before verification.
日本語の概要は準備中です。原文の説明を表示しています。
Tests a real web user flow with a browser by asserting semantic behavior, network and loading states, keyboard access, responsive layouts, and stable visual evidence. Use for browser bugs, end-to-end UI behavior, responsive or accessibility checks, and screenshot baselines. Not for static source review without a browser or for backend-only tests.
日本語の概要は準備中です。原文の説明を表示しています。