codex-first

Claude Code work routing: delegate implementation, fixing, exploratory subagents, rebasing, and PR merging/landing to GPT-6 Astra through Codex CLI while the parent specifies, decides, reviews, and verifies. Apply the native-Claude model gate. Codex-backed autoreview is always allowed and preferred.

Quality

81%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

88%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is highly actionable with executable launchers, explicit validation checkpoints, and feedback loops for fragile operations. Its only notable gap is progressive disclosure: the skill is a single large file that could offload the resume/watchdog detail into reference files.

Suggestions

Split the 'How resume works' and 'Liveness watchdog' sections into reference files (e.g. references/resume.md, references/watchdog.md), leaving a concise summary plus a one-level-deep pointer in SKILL.md.

Trim illustrative anecdotes (the minification incident) or move them into a references/case-studies.md to shorten the core body.

Add a brief table of contents or 'When to read which section' navigation aid at the top to improve discoverability across the long single-file layout.

Dimension	Reasoning	Score
Conciseness	The body is dense and high-signal — codex-cli flag/version specifics Claude would not already know — with no padding of basic concepts, though a few narrative anecdotes (e.g. the minification incident) and length could be trimmed.	4 / 5
Actionability	Copy-paste-ready, executable commands throughout: the `command codex exec --yolo -C <repo> -m gpt-6-astra …` launcher, the curl route probe, and the resume launcher cover the common cases concretely.	5 / 5
Workflow Clarity	Sequenced gate steps (1-2-3), explicit validation checkpoints ('check the log tail for `Usage: codex exec resume` before trusting the completion notification'), feedback loops (kill + resume same session), and a coordinator-verification checklist — including validation for the destructive rebase/merge/land operations.	5 / 5
Progressive Disclosure	Well-organized into clear sections (Hard gate, Route, Invoke, resume, watchdog, Verify) as a single-file skill with no bundle files, but large sub-topics like resume semantics and the liveness watchdog are inlined rather than split into one-level-deep reference files.	4 / 5
	Total	18 / 20 Passed

Description

75%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is specific and distinctive, listing concrete delegation actions and a clear Claude-vs-Codex division of labor. Its main weakness is the absence of an explicit 'Use when…' trigger clause, leaving the activation condition only weakly implied.

Suggestions

Add an explicit trigger clause, e.g. 'Use when the active session is Claude Code on a native Claude model and a hands-on task (implementation, fixing, rebase, PR landing) is worth delegating to Codex.'

Include natural synonyms a user might say ('hand off coding work', 'delegate to Codex', 'codex-first routing') to broaden trigger matching.

State the activation precondition in plainer terms so the 'when' is as concrete as the 'what'.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'delegate implementation, fixing, exploratory subagents, rebasing, and PR merging/landing' alongside 'specifies, decides, reviews, and verifies' — giving comprehensive coverage of the skill's capabilities.	5 / 5
Completeness	The 'what' is clear (delegate hands-on work to Codex CLI) but there is no explicit 'Use when…' trigger clause; the 'when' is only weakly implied by 'Apply the native-Claude model gate', which caps completeness at 3 per the missing-trigger guidance.	3 / 5
Trigger Term Quality	Domain-natural terms like 'Claude Code work routing', 'rebasing', 'PR merging/landing', 'Codex CLI', and 'autoreview' give good coverage, but natural synonyms and file extensions are absent.	4 / 5
Distinctiveness Conflict Risk	Occupies a clear niche — routing Claude Code hands-on work to Codex CLI / GPT-6 Astra — with distinct triggers and minimal overlap with other skills.	5 / 5
	Total	17 / 20 Passed

Validation

93%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
referenced_paths_exist	Referenced path issues: 3 missing	Warning

	Total	15 / 16 Passed

Repository: steipete/agent-scripts
Path: skills/codex-first/SKILL.md
Commit: 2784a38

Reviewed: 5 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.