Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add ytubecoder/codex-in-claude --skill peon-pokegit clone --depth 1 https://github.com/ytubecoder/codex-in-claudeWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/ytubecoder/codex-in-claude/peon-poke)<a href="https://agentmods.dev/skills/ytubecoder/codex-in-claude/peon-poke"><img src="https://agentmods.dev/badge/skills/ytubecoder/codex-in-claude/peon-poke.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00116 | $0.02979 |
| Opus 5 | $0.00058 | $0.01489 |
| Sonnet 5 | $0.00023 | $0.00596 |
| Haiku 4.5 | $0.00012 | $0.00298 |
Grade A, and why
peon-poke scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 62 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Peon Poke — Farm Implementation Work to External CLI Agents
/plan-check pulls opinions in; /peon-poke pushes labor out. An external agent (a "peon") does an implementation chore in an isolated git worktree; Claude — the foreman — reviews the work and merges it or sends it back. Nothing reaches the real working tree without review and an explicit merge.
All mechanics live in the peon script — never hand-roll worktrees or provider CLI calls. If command -v peon fails, tell the user to install it (README of ytubecoder/codex-in-claude) and stop.
Commands
| Command | Does |
|---|---|
peon dispatch <codex|grok|gemini|agy> "<task>" [--repo DIR] [--base REF] [--slug NAME] [--force] [--allow "glob[,glob...]"] [--verify "cmd"] |
Create worktree + peon/<slug> branch, run the provider inside it. Synchronous — background it to keep working. Fails loudly if the peon breaks contract (no commits / no report / dirty tree), preserving the worktree. --allow/--verify record the acceptance contract (file-scope allowlist + verify command) that check executes and merge enforces. |
peon list [--repo DIR] |
Active peons, report status, orphan detection |
peon report <slug> |
PEON_REPORT.md + commits + diffstat vs base + provider-self-reported token usage (codex/agy; n/a for grok/gemini) (+ dirty-tree warning). Usage is observability only — never quote it as "savings"; the counterfactual is unmeasurable. |
peon diff <slug> [--stat|--files] |
Diff vs base. --stat/--files for black-box review (never print the full diff you don't intend to read); bare = full diff for classic review |
peon check <slug> [--allow ...] [--verify ...] |
Mechanical acceptance gates, foreman-side: contract + allowlist scope + verify command run in the worktree (result recorded against the branch tip; a later poke invalidates it). Overrides are one-off when a contract was recorded at dispatch; when none was, they late-declare the contract (persisted, merge-enforced). |
peon adopt <slug> [-m "msg"] |
Foreman-commits work a sandbox-blocked provider left uncommitted (work + PEON_REPORT.md on disk, dispatch/poke exited 3). Refuses a clean worktree or a missing report. Adoption is mechanics, not acceptance — review still decides the merge. |
peon poke <slug> "<feedback>" |
Resume the same provider session for revisions; fails loudly on a no-op round |
peon merge <slug> [--into BRANCH] [--unchecked] |
After review: merge, strip the report, clean up. Refuses out-of-scope files and failed/stale/missing verify when a contract was recorded; --unchecked bypasses loudly. --into only asserts BRANCH is already checked out — peon never switches your branch. |
peon scrap <slug> |
Discard the work, remove worktree + branch + metadata |
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 8d ago First seen · 62 lines · 116 tokens per session scan A 45b05b135b39
peon-poke is a skill published in the GitHub repository ytubecoder/codex-in-claude (2 stars, last pushed 1mo ago), licensed MIT. It adds 116 tokens to every session and 2,979 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
cloudflare-workers-ci-cd
Complete CI/CD guide for Cloudflare Workers using GitHub Actions and GitLab CI. Use for automated testing, deployment pipelines, preview environments, secrets management, or encountering deployment failures, workflow errors, environment configuration issues.
claude-code-bash-patterns
Claude Code Bash tool patterns with hooks, automation, git workflows. Use for PreToolUse hooks, command chaining, CLI orchestration, custom commands, or encountering bash permissions, command failures, security guards, hook configurations.
definition-of-done
The ordered finish-line runner for any change in this repo — generators, then the asset-integrity refresh last on its own, then the full gate suite, then commit and push; use before declaring any change finished or whenever the stop hook reports uncommitted or untracked changes.
pr-babysit
The repeatable loop for watching a pull request until it is merged or closed, pairing webhook subscription with a scheduled self check-in so CI results, new pushes, and merge-conflict transitions are never missed; use when asked to watch, babysit, or autofix a PR, or right after creating one.
symlink-manager
Create, audit, repair, and document cross-platform symlinks that work correctly on both Windows and macOS/Linux. Use this skill whenever the user mentions symlinks, symbolic links, junction points, .gitconfig symlinks, broken links after git pull, cross-platform path issues, or needs help with ln -s equivalents on…
aws-ci-cd-release-engineer
Review AWS CI/CD and release safety across CodePipeline, CodeBuild, CodeDeploy, GitHub Actions, GitLab, artifact provenance, deployment gates, approvals, tests, progressive delivery, rollback, change correlation, and incident-prevention recommendations. Use when AWS releases or pipelines can affect production…