agent-bootstrap

Étoiles13

Forks2

Mis à jour19 mai 2026 à 13:15

Use when setting up or repairing repo agent instructions for Claude Code, Codex, or both. Proposes durable AGENTS.md, CLAUDE.md, and project memory changes before writing unless setup was explicitly requested.

Installation

Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.

Source

allytag

allytag/claude_code

Ouvrir le dépôt GitHub Voir les dépôts du créateur

Téléchargement

Métiers associésSOC

Basé sur la classification professionnelle SOC

Développeurs de logicielsProfessions informatiques et mathématiques·SOC 15-1252

SKILL.md

readonly

name	Agent Bootstrap
description	Use when setting up or repairing repo agent instructions for Claude Code, Codex, or both. Proposes durable AGENTS.md, CLAUDE.md, and project memory changes before writing unless setup was explicitly requested.
allowed-tools	Read, Bash, Edit, MultiEdit, Write

Agent Bootstrap

Use this skill to add durable project instructions without overwriting repo intent.

When To Use

User asks to bootstrap Claude Code, Codex, or AI agent workflow.
Repo lacks useful AGENTS.md / CLAUDE.md and the user wants setup or repair.
Existing instructions are stale, duplicated, unsafe, or too large.
Long project needs small boot files plus memory docs.

Do not use for normal coding unless instructions are part of the task.

Safety Rules

Read existing AGENTS.md, CLAUDE.md, README, package scripts, and relevant docs first.
Never overwrite existing instructions blindly.
Preserve project-specific rules; append guarded updates when possible.
Keep boot files short: commands, safety rules, memory location, verification expectations.
Put detailed context in docs/agent-memory/, not in boot files.
Do not add secrets, credentials, private paths, or one-project assumptions to reusable skills.
If boot files or docs/agent-memory/ are missing, do not auto-create them. Propose the files, summarize each in one line, and wait for explicit approval.
Exception: create directly only when the user explicitly asked for project setup, agent setup, memory setup, or said to set up the repo.

Boot File Shape

Include only:

project identity and critical flows
verified setup/test/build commands
destructive-action and secret rules
where memory docs live
when to update memory
final verification expectations

Workflow

Detect target: Claude Code (CLAUDE.md), Codex (AGENTS.md), or both.
Inspect existing files and commands.
Patch existing boot files when requested or clearly part of the task.
If files are missing, propose creating compact boot files and minimal memory index unless setup was explicitly requested.
Run diff review. Confirm no secrets or huge prompt dumps.

Output

Report files created/changed, preserved existing rules, and any assumptions not verified.

Plus depuis ce dépôt

même dépôt

api-contract

allytag/claude_code

Use for backend routes, API design, validation, database interactions, auth boundaries, response shapes, typed clients, error models, and integration-safe changes.

2026-05-1913

performance-pass

allytag/claude_code

Use for slow pages, large bundles, slow APIs, database performance, caching, rendering jank, long sessions, and production-readiness polish.

2026-05-1913

premium-ui

allytag/claude_code

Use for high-end UI, SaaS dashboards, landing pages, product websites, mobile responsive UI, visual redesign, UI polish, empty/loading/error states, motion, typography, and design quality rescue.

2026-05-1913

project-memory

allytag/claude_code

Use for long-running projects, context compaction, resume/handoff work, multi-session builds, or when the user asks to preserve project state. Proposes small verified memory docs before creating them unless setup was explicitly requested.

2026-05-1913

tool-coach

allytag/claude_code

Use automatically for Claude Code work involving files, tools, bash, edits, tests, long commands, failed tool calls, or unfamiliar repos. Improves tool choice, prevents loops, and keeps work high-quality without adding broad context.

2026-05-1913

code-review

allytag/claude_code

Senior code review workflow. Use when reviewing changes, checking quality before commit, evaluating generated code, or finding bugs/security risks.

2026-05-1913

Agent Bootstrap

Use this skill to add durable project instructions without overwriting repo intent.

When To Use

User asks to bootstrap Claude Code, Codex, or AI agent workflow.

Repo lacks useful AGENTS.md / CLAUDE.md and the user wants setup or repair.

Existing instructions are stale, duplicated, unsafe, or too large.

Long project needs small boot files plus memory docs.

Do not use for normal coding unless instructions are part of the task.

Safety Rules

Read existing AGENTS.md, CLAUDE.md, README, package scripts, and relevant docs first.

Never overwrite existing instructions blindly.

Preserve project-specific rules; append guarded updates when possible.

Keep boot files short: commands, safety rules, memory location, verification expectations.

Put detailed context in docs/agent-memory/, not in boot files.

Do not add secrets, credentials, private paths, or one-project assumptions to reusable skills.

If boot files or docs/agent-memory/ are missing, do not auto-create them. Propose the files, summarize each in one line, and wait for explicit approval.

Exception: create directly only when the user explicitly asked for project setup, agent setup, memory setup, or said to set up the repo.

Boot File Shape

Include only:

project identity and critical flows

verified setup/test/build commands

destructive-action and secret rules

where memory docs live

when to update memory

final verification expectations

Workflow

Detect target: Claude Code (CLAUDE.md), Codex (AGENTS.md), or both.

Inspect existing files and commands.

Patch existing boot files when requested or clearly part of the task.

If files are missing, propose creating compact boot files and minimal memory index unless setup was explicitly requested.

Run diff review. Confirm no secrets or huge prompt dumps.

Output

Report files created/changed, preserved existing rules, and any assumptions not verified.