本文へ移動
cccskills
無料GitHub で公開

resume

Continue dev work on worker tier. Use after thread rotation, turn limits, crashes, or fixing review/security findings. Reads the task file and resumes from the first unchecked task.

インストール方法を見る

含まれるファイル(1)

  • SKILL.md1.8 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。

Resume

Continue /architect work without chat history. Use a worker-tier model.

Follow the architect Phase 4–5 rules throughout (retry ceiling, self-review, thread rotation, completion message, implementation reflexes).

Start

  1. Read the task file in full — ## Session checkpoint first
  2. Rebuild the code context from the graph, not from chat history. bin/harness graft ask "<the area this task touches>" --source costs a few hundred tokens and returns exact file:line spans; reading the files it names to reconstruct that picture is the cost a rotation exists to avoid. Add bin/harness graft callers <symbol> --depth 2 when the remaining tasks change a signature. The rule and the full command table are in the graft skill.
  3. If the file does not exist or cannot be read, say: "Task file not found at <path>. Please provide the correct path or describe what we were working on."
  4. Say: "Resuming from task N. Remaining: [list]. Next: [action]."
  5. Execute from the first unchecked task (architect Phase 4–5 rules)

Reference docs

Grep one section only — never load a whole .agents/docs/ file. They run to megabytes and one section is almost always the answer.

Plan changes

Update ## Tasks and note in ## Log if the remaining plan is wrong.

Hand-off

Follow review_required and lint tiers from the task file and AGENTS.md.

Project specifics

Paths, commands and build steps for this project are in AGENTS.md and workflow.md. Read them rather than assuming the layout of another project in this fleet.

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

architect

無料

Full dev cycle. Planning tier for clarify and plan; worker tier after approval for autonomous execution, tiered lint, thread rotation, and feedback fixes. Start every change here.

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。

quentin-ld/zenpress92026年10月9日 更新

Writing documents an agent consumes. Use when creating or editing a skill, an AGENTS.md, a README section, a docblock, or any file reached by a pointer.

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。

quentin-ld/zenpress92026年10月9日 更新

graft

無料

This repo is indexed by graft/. For ANY task here, whether understanding how something works, finding where code lives, tracing what calls a symbol or what a change breaks, or scoping an edit, get your context from the graph before grepping or reading source files. Every command runs through `bin/harness graft`.

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。

quentin-ld/zenpress92026年10月9日 更新

grilling

無料

Interviewing the owner to a shared understanding before any work starts. Use when a plan, a decision or an idea needs stress-testing, or when the owner asks to be grilled, interviewed or questioned about one.

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。

quentin-ld/zenpress92026年10月9日 更新

Removing AI tells from French prose without lowering its register. Use when rewriting or reviewing French text — a reply to the owner, a `.po` target, a French post — that reads like a machine wrote it.

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。

quentin-ld/zenpress92026年10月9日 更新

humanizer

無料

Editing English prose that reads as machine-written. Use when reviewing or rewriting a draft for AI tells — a README, a changelog entry, a comment, a user-facing string, a review. French text is `humaniseur-fr`.

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。

quentin-ld/zenpress92026年10月9日 更新

quentin-ld のスキルをすべて見る

このスキルの問題を報告する