Audit or rewrite AGENTS.md so it holds only lasting principles. Run it only when the user asks for it by name; never invoke it on your own.
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。
dzhng/skills1,0262026年10月6日 更新
無料GitHub で公開
implement-spec-with-codex
Run [implement-spec](../implement-spec/SKILL.md) with Codex doing the implementation passes while you orchestrate, integrate, and review. Use when the user asks to implement a spec with Codex, or to run implement-spec with Codex as the implementer.
インストール方法を見る含まれるファイル（1）
- SKILL.md2.1 KB
SKILL.md（原文）
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。
Implement Spec with Codex
Run implement-spec exactly — same loop, same
gates, same maintenance checkpoints — but Codex writes the code. Invoking this
skill IS the explicit ask that authorizes codex exec delegation under the
codex skill.
Role split
- Codex implements — whenever possible. Default every implementation pass
to a
codex exec, sliced sharp and prompted per the codex skill (prompt blocks, sandbox choice, exec liveness). Run independent slices as concurrent execs in separate worktrees, exactly where implement-spec would fan out subagents. - You orchestrate and review. The spec loop stays yours: pick the next slice, reconcile plan with code, author each prompt end-to-end, integrate diffs, run the gates, review the merged tree, commit, update the handoff, run maintenance checkpoints. None of that delegates.
Keep a pass yourself only when slicing it sharply would cost more than doing it — tiny fixups, plan reconciliation, integration conflicts.
Rules
- You own every result: read the full diff and run the verification yourself. "Codex says it's done" is never done.
- Visual results carry implement-spec's two extra proofs (production-route pixel diff + unprimed critique) with double force here: a sandboxed Codex never saw its change render, so a pass that misses the production render path entirely arrives looking fully "verified".
- A sandboxed Codex ships code it never saw run — browser-verify every visual slice yourself and budget fix rounds. Send a red gate back as a resumed follow-up with the failing evidence; fix it yourself only when that's smaller than re-prompting.
- Commit hygiene stays yours: review the change list of the merged tree before every commit, exactly as implement-spec demands of the integrating reviewer.
レビュー
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。
同じリポジトリのスキル
概要と使いどころ
Audit the choices an implementing agent made, not its diff — a pure decision audit that traces the session's history into a choices ledger, changes no code, and never blocks an unsupervised run. Working code still embeds architecture the user never chose; surface it because future work inherits it. Use when the user wants to review the decisions the AI made on their behalf, before merging or committing AI-implemented work, when integrating a delegated subagent's pass, or when a fix "works" but might be a point fix.
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。
dzhng/skills1,0262026年10月6日 更新
Audit and prioritize performance work through bounded-work and forward-progress checks. Use when asked to find performance issues, investigate freezes/500s/OOMs, review retries or polling for no-progress loops, rank a performance backlog, or implement the next simple performance fixes.
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。
dzhng/skills1,0262026年10月6日 更新
Audit whether tests earn their maintenance cost and which suite owns each contract. Use when pruning redundant tests, investigating implementation coupling or test-only production hooks, reviewing the value of proposed coverage, or auditing an entire subsystem. Use write-tests to implement the resulting test changes.
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。
dzhng/skills1,0262026年10月6日 更新
Run fast, progressive experiments to map tunable parameters and their effects. Use when optimizing code, prompts, configurations, or other artifacts against a goal or benchmark, or when a long research run needs a planned hypothesis queue, focused trials, and durable findings.
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。
dzhng/skills1,0262026年10月6日 更新
claude
無料Use Claude Code as an independent `claude -p` subagent when the user explicitly asks for Claude, wants a second-agent opinion from Claude, or asks to delegate a well-scoped task to Claude. Supports selecting `--model` and thinking/effort level with defaults of `opus` and `high`.
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。
dzhng/skills1,0262026年10月6日 更新