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 agentmods add skills/congmnguyen/claude-code-wsl2-setup/codex-delegatenpx skills add congmnguyen/claude-code-wsl2-setup --skill codex-delegategit clone --depth 1 https://github.com/congmnguyen/claude-code-wsl2-setupWrote 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/congmnguyen/claude-code-wsl2-setup/codex-delegate)<a href="https://agentmods.dev/skills/congmnguyen/claude-code-wsl2-setup/codex-delegate"><img src="https://agentmods.dev/badge/skills/congmnguyen/claude-code-wsl2-setup/codex-delegate.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 | $0.00109 | $0.01057 |
| Opus 5 | $0.00055 | $0.00528 |
| Sonnet 5 | $0.00022 | $0.00211 |
| Haiku 4.5 | $0.00011 | $0.00106 |
Grade A, and why
codex-delegate 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 4d 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 — 52 lines — stays where its author put it; the contents beside it link to each section on GitHub.
What this does
Spawns the codex-delegate subagent (low-effort Sonnet), which runs codex exec with output redirected to a log and returns only a ~10-line summary. Routing lives in ~/.claude/CLAUDE.md § Codex Delegation; wrapper mechanics live in the agent definition. This file covers only the orchestrator side: how to spawn and what to do with the result.
How to invoke
-
Turn the request (
$ARGUMENTSif given, else the current task) into a self-contained spec — Codex has no memory of this conversation, so include everything: exact behavior, file paths, edge cases, and how to verify. -
Spawn the subagent (background), passing TASK, WORKDIR (default: current working directory), SANDBOX (
workspace-writefor edits,read-onlyfor investigation), and a VERIFY command if there's a test/build to check.Preferred — the dedicated agent (available after a Claude Code restart that registers
.claude/agents/codex-delegate.md):Agent( subagent_type: "codex-delegate", description: "delegate to codex", prompt: "TASK: <full self-contained spec>\nWORKDIR: <abs path>\nSANDBOX: workspace-write\nVERIFY: <test/build cmd or omit>\nTAKEOVER: allowed # optional — lets the wrapper finish mechanical remainder itself if Codex is quota-blocked/crashed; omit to restrict it to salvage-only", run_in_background: true )Fallback — if that errors with
Agent type 'codex-delegate' not found(the definition isn't loaded in this session yet), spawn a generic Sonnet subagent that reads the same playbook file:Agent( subagent_type: "general-purpose", model: "sonnet", effort: "low", description: "delegate to codex", prompt: "First Read /home/cong/.claude/agents/codex-delegate.md and adopt it as your complete role and instructions. Then carry out:\nTASK: <full self-contained spec>\nWORKDIR: <abs path>\nSANDBOX: workspace-write\nVERIFY: <test/build cmd or omit>", run_in_background: true )
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.
- 4d ago First seen · 52 lines · 109 tokens per session scan A 62a4ca8137f0
codex-delegate is a skill published in the GitHub repository congmnguyen/claude-code-wsl2-setup (48 stars, last pushed 3d ago), licensed MIT. It adds 109 tokens to every session and 1,057 once invoked, about $0.0005 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 2026-08-30.
Other skills, from other repositories
yt-dlp-music-downloads
Download YouTube playlists into the Navidrome music library.
add-system-action
Use when adding a new TUI button/action (system action) or a new package-manager backend to personal-os-setup — e.g. "add a button to do X", "support a new package manager", "add a new tab/section". Covers the factory.py section-builder pattern, the managers/shared.py boilerplate pattern, and the required test…
run-tests
Use when writing or running tests in this repo, or when a test fails unexpectedly — "add a test", "run the tests", "why did this test fail/get skipped", "test the new manager/action". Covers make test vs make test-integration (the latter is destructive against the host!), the three test suites, and the exact…
docs-site
Use when adding/editing documentation pages, or building/previewing/deploying the properdocs (mkdocs) site — "add a doc page", "preview the docs site", "deploy docs", "why isn't my new page showing up". Covers properdocs.yml's docsdir:. quirk, the exclude-glob trap for new top-level directories, and the auto-generated…
fork-and-customize
Use when someone wants to fork this repo to build their own personal OS setup tool, or asks how to rebrand/adapt it — "fork this repo", "make this my own", "use this for my dotfiles", "rebrand this". Covers what to rename/rebrand, how to customize packages.yaml and the chezmoi dotfiles source, and what release/CI…
repo-gotchas
Use before trusting CONTRIBUTING.md/Makefile/pre-commit claims literally, or when something documented doesn't behave as expected — "why doesn't X work", "is this hook actually active", "make help is missing stuff". Documents known drift between this repo's docs and its actual behavior.