Prepares and controls software releases with risk classification, CI quality gates, artifact provenance, feature flags, staged rollout, measurable promotion and rollback thresholds, post-deploy verification, and cleanup. Use for deployment pipelines, launch readiness, canaries, production rollout, release automation, or rollback planning. Not for merely merging a completed development branch or proving one local change.
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。
Turn a release into a sequence of evidence-backed decisions rather than one irreversible event.
Record the artifact/version, environments, owner, change scope, affected users, data or contract changes, dependencies, risk class, maintenance window, and success metrics. Use release-record.md for the go/no-go and rollout record.
Require checks proportional to risk: build and tests, contract and migration proof, security findings, performance budgets, smoke paths, accessibility or visual checks, artifact signing/provenance, monitoring, and a tested recovery path. A green generic pipeline is not proof of the release's critical behavior.
For canaries and production observation windows, use
post-deploy-verification.md. Confirm
the deployed build marker, run critical browser journeys through
browser-testing, and record each promote, hold, or rollback decision against
fresh baseline evidence.
Use immutable artifacts across environments. Keep release credentials out of logs and command lines. Require explicit authority for production mutation.
Verify critical paths after deployment, monitor delayed jobs and regional
effects, remove stale flags and compatibility paths on schedule, update release
notes, and create owned follow-ups for non-blocking debt. Never substitute a
successful deployment command for an observation window. Use
incident-response if live impact begins and safe-migrations for stateful
cutovers.
The intended artifact reached the intended audience, promotion evidence and remaining limits are recorded, rollback readiness was real, and temporary release controls have owners and expiry dates.
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。