菜单
Create, update, and maintain AGENTS.md and agents-docs/ for best practices. Handles documentation structure, skill creation, and reference management.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Enhance session checkpoints with git context and diff summaries. Use after completing significant work to enrich checkpoint.md files with repository state, recent commits, and code change summaries. Triggers include "enhance checkpoint", "update checkpoint with git", "add git context to checkpoint", or when finishing a multi-file change session.
Use for model training, hyperparameter tuning, and Modal GPU training.
Use when invoking training, evaluation, and dataset preparation via CLI. Provides all standard commands for this project.
Multi-perspective code analysis using three AI personas (RYAN, FLASH, SOCRATES) for comprehensive decision-making. Use when complex code decisions need analysis from multiple viewpoints, or when avoiding single-perspective blind spots is critical.
Use for authentication management, token validation, and credential troubleshooting.
Use for secrets management, credentials handling, and security best practices.
| name | agents-md |
| description | Create, update, and maintain AGENTS.md and agents-docs/ for best practices. Handles documentation structure, skill creation, and reference management. |
Manage AGENTS.md and related documentation following project conventions.
Keep AGENTS.md under 120 lines. Include:
Extended docs in separate files:
agents-docs/
├── training.md # GPU, Modal, options
├── ci-cd.md # Fix workflow, quality gate
├── security.md # Secrets, tokens
├── skills.md # Agent specialization
└── learnings.md # Patterns, ADRs
Follow .agents/skills/<skill-name>/SKILL.md structure:
---
name: <skill-name>
description: <1-line description>
---
# Skill: <skill-name>
<detailed content under 250 LOC>
Requirements:
When adding new features:
bash scripts/quality-gate.shAll skills must pass validation:
bash scripts/quality-gate.sh
Checks: