review-plan

スター9

フォーク1

更新日2026年7月22日 14:08

Explicit-only. Review and repair a plan/task/context artifact set with independent evidence, then report implementation readiness.

インストール

Codex または Claude でインストールこの Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。

ソース

conn-castle

conn-castle/agent-layer

GitHub リポジトリを開く Creator のリポジトリを見る

ダウンロード

review-plan

Required inputs

Require:

plan, task, and context artifact paths
an optional specification artifact path

Findings artifact

Write .agent-layer/tmp/review-plan.<run-id>.findings.md with run ID YYYYMMDD-HHMMSS-<short-rand>.

Workflow

Do not edit the input artifacts or implementation code.

Review all inputs directly through 1–3 distinct framings chosen for the plan's risk and complexity. Each framing covers the complete artifact set; do not split coverage. Consequential architecture changes require an architecture framing.
Assess every potential finding against the artifacts and repository. Do not treat repetition across framings as evidence.
Report only evidence-backed findings that are material. For each, give its location, evidence, impact, and suggested correction.

Include the framing count, rationale, framings, and any substantive user decision. Return the findings path and one verdict:

approved: no material findings
changes-needed: material findings remain
blocked-for-user-decision: a substantive choice is unresolved

このリポジトリの他の Skills

同じリポジトリ

auto-skill-loop

conn-castle/agent-layer

Explicit-only. Run a named autonomous mode, repeatedly selecting, implementing, and shipping coherent work while preserving anything blocked.

2026-08-019

playwright

conn-castle/agent-layer

Use globally installed `playwright-cli` for browser automation, UI inspection, screenshots, Playwright test debugging, or test generation. Trigger when a task needs browser interaction or test repair. Do not use for generic tests, web search, non-browser docs, or API-only work.

2026-07-319

full-workflow

conn-castle/agent-layer

Explicit-only. Align a feature specification, produce a reviewed plan, complete the local work, and ship the pull request.

2026-07-319

ship-pr

conn-castle/agent-layer

Explicit-only. Ship one pull request, observe hosted CI and review feedback, follow PR policy, reply to every eligible comment, and request merge authorization before merging and cleaning up.

2026-07-319

verify-work

conn-castle/agent-layer

Explicit-only. Verify completed work against its authoritative contract and report coverage, working-code evidence, material gaps, and the completion verdict.

2026-07-319

agent-dispatch

conn-castle/agent-layer

Use Agent Dispatch MCP tools only when the user names an external dispatch target or another skill explicitly requires dispatch. Do not use it for generic subagent, second-agent, or fresh-context requests; use the built-in subagent instead.

2026-07-319

name	review-plan
description	Explicit-only. Review a plan/task/context artifact set and report material findings.

review-plan

Required inputs

Require:

plan, task, and context artifact paths
an optional specification artifact path

Findings artifact

Write .agent-layer/tmp/review-plan.<run-id>.findings.md with run ID YYYYMMDD-HHMMSS-<short-rand>.

Workflow

Do not edit the input artifacts or implementation code.

Review all inputs directly through 1–3 distinct framings chosen for the plan's risk and complexity. Each framing covers the complete artifact set; do not split coverage. Consequential architecture changes require an architecture framing.
Assess every potential finding against the artifacts and repository. Do not treat repetition across framings as evidence.
Report only evidence-backed findings that are material. For each, give its location, evidence, impact, and suggested correction.

Include the framing count, rationale, framings, and any substantive user decision. Return the findings path and one verdict:

approved: no material findings
changes-needed: material findings remain
blocked-for-user-decision: a substantive choice is unresolved