RonanCodes/codex
skills/codex/SKILL.md
One-shot headless OpenAI Codex CLI calls for background/admin AI tasks — summaries, classification, extraction, admin glue. The default engine for anything that runs AI constantly in the background (daemon-driven, per-event), because it bills the flat ChatGPT subscription instead of Claude usage or per-token API spend, and it keeps working while Claude is rate-limited. NEVER for coding — coding stays Claude. Use when a skill or daemon needs a cheap always-on AI call, when the user says "use codex", "ask codex", "codex as backup", or when building a background summarizer/classifier into a listener or loop. Reads auth from ~/.codex/auth.json (ChatGPT account, no API key).
tools
Updated Jul 5, 2026
$ install --global
skillsauthnpx skillsauth add RonanCodes/ronan-skills codexInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: Jul 5, 2026, 2:32 AM117.5s3 files scanned
SKILL.md
- name:
- codex
- description:
- One-shot headless OpenAI Codex CLI calls for background/admin AI tasks — summaries, classification, extraction, admin glue. The default engine for anything that runs AI constantly in the background (daemon-driven, per-event), because it bills the flat ChatGPT subscription instead of Claude usage or per-token API spend, and it keeps working while Claude is rate-limited. NEVER for coding — coding stays Claude. Use when a skill or daemon needs a cheap always-on AI call, when the user says "use codex", "ask codex", "codex as backup", or when building a background summarizer/classifier into a listener or loop. Reads auth from ~/.codex/auth.json (ChatGPT account, no API key).
- category:
- workflow
- argument-hint:
- <prompt> [--timeout <seconds>] [--model <model>]
- allowed-tools:
- Bash(bash *) Bash(python3 *) Bash(codex *) Read
Codex — background/admin AI engine
One-shot, headless calls to the OpenAI Codex CLI. This is the canonical implementation of canon/codex-for-background-ai.md: always-on background AI defaults to Codex; Codex never codes.
When to reach for this
- A daemon or loop needs an AI call on every event (run summaries, log triage, one-line classification) — that must not burn Claude usage, and it must keep working when Claude itself is rate-limited.
- A generic admin AI task where the answer is text, not code changes.
Do NOT use it to write, review, or refactor code, or as a judge of whether Claude is back after a rate limit (that stays deterministic — see /ro:linear-listen § "Usage-limit recovery").
How to call it
From shell (or another skill):
bash ~/Dev/ronan-skills/skills/codex/scripts/ask.sh "Summarize in one sentence: <text>"
# stdout = the reply · exit 3 = codex unavailable/failed → fall back
From Python (what linear-listen does):
import importlib.util
p = Path("~/Dev/ronan-skills/skills/codex/scripts/codex_ask.py").expanduser()
spec = importlib.util.spec_from_file_location("codex_ask", p)
codex_ask = importlib.util.module_from_spec(spec); spec.loader.exec_module(codex_ask)
reply = codex_ask.ask(prompt) # str, or None → fall back
ask() never raises: missing binary, timeout, non-zero exit, or empty reply all return None, so the caller's fallback chain stays trivial.
The invocation it makes
codex exec --skip-git-repo-check -s read-only -o <tmpfile> [-m <model>] "<prompt>"
-s read-only— background tasks get no write/exec sandbox.-o(--output-last-message) — the reply is read from the file; stdout is a banner, never parsed.- Runs with
cwd=$HOMEso no repo context is loaded. - Timeout 180s by default (
CODEX_ASK_TIMEOUTor--timeout); a trivial prompt answers in ~7s. - Model: whatever the Codex CLI defaults to (gpt-5.5 as of 2026-07-04); override with
--modelonly when needed.
The fallback chain (callers implement this)
Codex CLI → OpenAI API (OPENAI_API_KEY) → deterministic heuristic. Every layer optional, the last one never fails. First consumer: skills/linear-listen/scripts/server.py summarize_run.
Auth / setup
The Codex CLI caches the ChatGPT-account token in ~/.codex/auth.json (or the OS keyring) after a one-time interactive codex login. No env key. If auth is missing the call fails → None → the fallback chain handles it; run codex login in a terminal to fix.
Note: ChatGPT-account Codex only supports the models the plan exposes (e.g. gpt-5.3-codex is rejected; the default works). Don't hardcode a model.
Related
canon/codex-for-background-ai.md— the rule and rationale/ro:linear-listen— first consumer (run-outcome summaries) + the deterministic rate-limit recovery this backs up- LLM wiki:
llm-wiki-skill-lab/wiki/patterns/codex-for-background-ai.md
Related Skills
RonanCodes/skills/linear-pipeline
testing
VerifiedTrustedCommunity
--- name: linear-pipeline description: The Fable orchestrator for a single dispatched Linear ticket. Holds almost no context itself; it receives `--issue <ID> --detached`, decides the stage sequence, and fans out a sub-agent per stage, passing forward only each stage's artifact (never re-derived, never inlined into its own context). Step zero, before any planning or stage routing, is a boundary triage against `canon/security-boundary.md` (#199): a match tags Ronan Connolly and stops the run, no
SKILL.mdUpdated Jul 5, 2026
RonanCodes/skills/in-your-face
development
VerifiedTrustedCommunity
--- name: in-your-face description: Capture a chat-only answer into a durable artifact (markdown + HTML, PDF when cheap) and launch it automatically so the user cannot miss it. Use when user says "in your face", "don't let me lose this", "save that answer", "make that durable", or right after answering a substantive side question (a recipe, comparison, how-to, or generated prompt) that would otherwise die with the context. category: workflow argument-hint: [--no-open] [--vault <short>] [hint of
SKILL.mdUpdated Jul 5, 2026
RonanCodes/warranty-check
research
VerifiedTrustedCommunity
Turn a warranty rejection, repair quote, or RMA email into a cited decision brief — legal read (NL/EU consumer law), is the part user-serviceable, live part and new-unit prices, repair-vs-DIY-vs-new economics, before-you-send-it checklist, deadlines. Use when the user pastes or screenshots a repair quote, warranty rejection, "not covered" email, onderzoekskosten fee, or asks "should I repair or replace this".
SKILL.mdUpdated Jul 4, 2026
RonanCodes/boardgamegeek
development
VerifiedTrustedCommunity
Interact with the BoardGameGeek XML API2. Search games, fetch details with ratings and weight, pull a user's collection or wishlist, download box art from the CDN. Covers the 2026 auth reality (registered application token, Bearer header, 401 without one, approval takes a week or more) plus a browser fallback via the GEEK.geekitemPreload JSON while approval is pending. Reads BGG_XML_API_TOKEN from ~/.claude/.env. Triggers on "boardgamegeek", "bgg", "fetch my BGG collection", "look up a board game", "bgg search", "download board game images".
SKILL.mdUpdated Jul 4, 2026