Search skills.../

Atft Autonomy | Skills Pool

Skill File

Atft Autonomy

Coordinate Claude Code skills with OpenAI Codex autonomous workflows for end-to-end ATFT-GAT-FAN maintenance.

wer-inc0 starsOct 23, 2025

Occupation: Network and Computer Systems Administrators
Categories: CI/CD

Skill Content

ATFT Autonomy Skill

Mission

Seamlessly orchestrate Claude and Codex agents to keep the ATFT-GAT-FAN stack production-ready.
Decide when to run deep Codex analyses versus fast Claude remediations.
Maintain shared context (AGENTS.md, health snapshots, run logs) across agents.

Engagement Signals

Requests to “run full autonomous maintenance”, “coordinate Claude and Codex”, or “schedule daily self-healing”.
Situations where an issue spans multiple domains (dataset + training + quality).
Need to generate or refresh .mcp.json, AGENTS.md, or other shared configs.

Preflight Checklist

Confirm tool availability:
- command -v codex (expect npm package @openai/codex).
- command -v ./tools/claude-code.sh.

Related Skills

Quick Install

Atft Autonomy

npx skills add wer-inc/gogooku3

Download Skill Open repository

Author: wer-inc
stars: 0
Updated: Oct 23, 2025
Occupation

On this page

01ATFT Autonomy Skill

Verify configuration files:

.mcp.json exists with filesystem/git servers (tools/codex.sh recreates if missing).
AGENTS.md up to date with latest guidelines.

Run snapshot diagnostics:

tools/project-health-check.sh --summary (logs to _logs/health-checks/).
nvidia-smi and df -h to ensure resource headroom before long autonomous jobs.

Core Playbooks

1. Full Daily Autonomy Loop

./tools/claude-code.sh --no-check "Run proactive maintenance checklist" — quick fixes + TODO updates.
./tools/codex.sh --max --exec "Perform deep optimization sweep across dataset, training, and research modules" — longer Codex session with MCP.
Append combined findings to docs/ops/autonomy_log.md (capture commands run, deltas, next steps).
git status --short → review diffs, ensure no unintended drift.

2. Incident Response (multi-domain failure)

Launch Claude for rapid triage: ./tools/claude-code.sh "Investigate failed training and prepare briefing for Codex".
Pass Claude findings to Codex:
./tools/codex.sh --exec "Read docs/ops/incident_brief.md and propose remedial plan".
Execute agreed actions (dataset rebuild, training rerun, quality checks).
Update docs/ops/incident_log.md with both agents’ actions and resolutions.

3. Scheduled Autonomous Optimization

Add cron entry (example):
0 2 * * * cd /workspace/gogooku3 && ./tools/codex.sh --max >> _logs/daily-optimization.log 2>&1.
Pair with weekly Claude sweep:
0 7 * * MON cd /workspace/gogooku3 && ./tools/claude-code.sh --no-check >> _logs/weekly-claude.log 2>&1.
Summarize improvements weekly by aggregating _logs/codex-sessions/*.log and _logs/claude-code/*.log; record highlights in docs/ops/weekly_autonomy_report.md.

Shared Context Management

Ensure both agents read CLAUDE.md and AGENTS.md before editing high-risk files.
Rotate docs/SKILLS_GUIDE.md and claude/skills/ when procedures change.
Store cross-agent decisions in EXPERIMENT_STATUS.md or docs/ops/autonomy_log.md.

Failure Handling

Codex CLI missing → npm install -g @openai/codex.
MCP server errors → regenerate .mcp.json via ./tools/codex.sh and restart run.
Conflicting edits → consolidate drafts under docs/ops/autonomy_pending/ and resolve manually.
Resource saturation → stagger Claude/Codex runs, cap concurrency in cron.

Handoff

After combined sessions, notify stakeholders by updating docs/ops/weekly_autonomy_report.md.
Attach session logs _logs/codex-sessions/*.log and _logs/claude-code/*.log for audit.
Re-sync ~/.claude/skills/ so Claude picks up any procedural updates discovered during Codex runs.

02

Mission

03Engagement Signals

04Preflight Checklist

05Core Playbooks

061. Full Daily Autonomy Loop

072. Incident Response (multi-domain failure)

083. Scheduled Autonomous Optimization

09Shared Context Management

10Failure Handling

Feishu Wiki

Feishu knowledge base navigation. Activate when user mentions knowledge base, wiki, or wiki links.

Automation Tools

Coding Agent (bash-first)

Delegate coding tasks to Codex, Claude Code, or Pi agents via background process. Use when: (1) building/creating new features or apps, (2) reviewing PRs (spawn in temp dir), (3) refactoring large codebases, (4) iterative coding that needs file exploration. NOT for: simple one-liner fixes (just edit), reading code (use read tool), thread-bound ACP harness requests in chat (for example spawn/run Codex or Claude Code in a Discord thread; use sessions_spawn with runtime:"acp"), or any work in ~/clawd workspace (never spawn agents here). Claude Code: use --print --permission-mode bypassPermissions (no PTY). Codex/Pi/OpenCode: pty:true required.

CMS & Platforms

Prose

OpenProse VM skill pack. Activate on any `prose` command, .prose files, or OpenProse mentions; orchestrates multi-agent workflows.

Clawhub

Use the ClawHub CLI to search, install, update, and publish agent skills from clawhub.com. Use when you need to fetch new skills on the fly, sync installed skills to latest or a specific version, or publish new/updated skill folders with the npm-installed clawhub CLI.

System Administration

Openhue

Control Philips Hue lights and scenes via the OpenHue CLI.

Openclaw Parallels Smoke

End-to-end Parallels smoke, upgrade, and rerun workflow for OpenClaw across macOS, Windows, and Linux guests. Use when Codex needs to run, rerun, debug, or interpret VM-based install, onboarding, gateway smoke tests, latest-release-to-main upgrade checks, fresh snapshot retests, or optional Discord roundtrip verification under Parallels.

Network and Computer Systems Administrators