Borrowing it
Nothing to install: this file belongs to luthersystems/elps. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/luthersystems/elps/main/.claude/skills/codex-delegate/SKILL.mdgit clone --depth 1 https://github.com/luthersystems/elpsWrote 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/luthersystems/elps/codex-delegate)<a href="https://agentmods.dev/skills/luthersystems/elps/codex-delegate"><img src="https://agentmods.dev/badge/skills/luthersystems/elps/codex-delegate/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/luthersystems/elps/codex-delegate"><img src="https://agentmods.dev/badge/skills/luthersystems/elps/codex-delegate.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00000 | $0.03136 |
| Opus 5.5 | $0.00000 | $0.01254 |
| Sonnet 5.5 | $0.00000 | $0.00627 |
| Haiku 4.5 | $0.00000 | $0.00314 |
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 2d 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 — 255 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/codex-delegate — Codex Delegation Skill
Hands a bounded coding unit to OpenAI Codex in its own git worktree, with write access to the files, and takes the work back verified on the host.
Trigger
Use when asked to delegate work to Codex, to get a second opinion or an independent diagnosis from Codex, to fan out several Codex tasks at the same time, or when a Codex run says the broker is busy. Also covers the set-up this container needs, because the Codex login and the plugin do not survive a new session.
Use scripts/codex-delegate.sh. It calls the Codex plugin's own runtime
(codex-companion.mjs). It does not copy the plugin logic.
WARNING: the working tree may be the only copy
Codex CANNOT COMMIT in this sandbox: .git is read-only. Git cannot recover
unstaged edits destroyed by git checkout HEAD -- ., git restore,
git reset --hard, git clean, or git worktree remove --force. Never tidy
an active or uncollected unit's worktree. A transcript is not a backup.
The wrapper captures before launch, every second while monitoring, and before
and after teardown. Each run's current symlink points to a complete immutable
snapshots/snapshot-* directory. tracked.patch includes staged and unstaged
changes and binary payloads relative to base.commit; untracked/ preserves
nonignored new files and symlinks. untracked.paths0 is a NUL-delimited manifest.
Earlier snapshots are retained even if a later command removes the changes.
A failed capture leaves the previous snapshot available and prints an error.
To recover, select a snapshot containing the work, create a separate clean
worktree at the commit in its base.commit, apply its tracked.patch with
git apply, then copy untracked/. into that worktree with cp -a. Inspect
and commit the recovered work from an environment that can write .git.
After SIGKILL, exit-status and the final message may be absent: look in
snapshots/ and in runs/recovery.* before touching the source worktree.
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.
- 2d ago First seen · 255 lines · 0 tokens per session scan A 2de36d66a075
codex-delegate is a skill published in the GitHub repository luthersystems/elps (36 stars, last pushed today), licensed BSD-3-Clause. It costs nothing until one of its globs matches a file; then it loads 3,136 tokens. 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 2026-10-26.
Other skills, from other repositories
vhs-e2e-gif
Record a test run, a TUI session, or any terminal command as a GIF with VHS and attach it to a GitHub PR as a release-hosted asset, never a repo commit. Use when asked to record an e2e run, demo a fix on a PR, attach a GIF or screen recording to a pull request, show a test passing visually, or produce a terminal…
check-linters-before-commit
Before any commit, run linters, vet, tests, and build verification; do not commit until checks pass.
gh-labels
Reference for what GitHub issue/PR labels mean in this workspace (S-triage, S-needs-scope, S-impl, prio-p0-critical, A-ext-dev-tools, A-perf, A-ux, and others under My Labels). Use when the user asks what a label means, which label to apply, or how to add missing labels to a repo. Secondary: bootstrap labels on a new…
gh-pr
Writes clear pull request descriptions from the current branch diff against a base branch. Defaults to main, but supports dev or any other user-specified comparison branch. Use when the user asks for a PR description, pull request summary, or a markdown write-up for changes against a base branch.
gh-rev
Use the gh-rev CLI to maintain a persistent Markdown review ledger for local branches and pull requests. Trigger when the user names gh-rev, asks to save or submit a durable review, inspect or implement existing findings, continue an independent review pass, or adopt branch history into a PR. Use the gh-pr-review…
gh-ci-watch
Watch GitHub PR checks with gh, poll until green/failure/timeout, fetch failed GitHub Actions logs, save logs locally, summarize actionable CI failures, and apply trivial local fixes when safe. Use when the user asks to watch PR checks, wait for CI, inspect failed actions, fix obvious CI failures, or summarize gh pr…