codex

Delegate coding to OpenAI Codex CLI (features, PRs).

Quality

64%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Medium

Suggest reviewing before use

Fix and improve this skill with Tessl

tessl review fix ./skills/autonomous-ai-agents/codex/SKILL.md

Quality

Content

76%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 copy-paste-ready commands across many workflows and is well-organized and mostly lean. Its main weakness is workflow clarity: batch and destructive workflows lack integrated validation checkpoints, with safety guidance relegated to a caveat section instead of being embedded as explicit steps.

Suggestions

Embed explicit validation checkpoints in the batch/destructive workflows (e.g. 'git diff review', 'run targeted tests', 'confirm before push') as numbered steps rather than only mentioning them in the Hermes Gateway Caveat.

Add a verify step after each Codex run (e.g. review the generated diff and run tests) before the push-and-create-PR step in the worktree and batch-review workflows.

Tighten the auth-detail paragraph around ~/.codex/auth.json to remove redundant explanation.

Dimension	Reasoning	Score
Conciseness	Dense and practical with no padding of concepts Claude already knows; minor over-explanation in the auth-detail paragraph and the lengthy (though justified) Hermes gateway caveat could be trimmed. Not a 5 because a few sections can be tightened.	4 / 5
Actionability	Abundant copy-paste-ready terminal() and process() commands covering one-shot, background, PR review, worktree, and batch-review cases; all are real executable shell with no pseudocode. Matches the fully executable, common-case-covering anchor.	5 / 5
Workflow Clarity	Sequences are clear, but the batch/destructive workflows (worktree push→PR, batch PR reviews) lack integrated validation checkpoints — the diff-review/test/confirmation guidance lives only in the caveat section. Per the rubric, missing validation in batch/destructive operations caps this at 3. Not a 4 because checkpoints are absent from the workflows themselves.	3 / 5
Progressive Disclosure	No bundle files exist and all content is inline, but it is well-organized into clearly headed sections; at ~140 lines the under-50-line 5-exception does not apply. Not a 5 because there are no signaled one-level-deep references or content splits.	4 / 5
	Total	16 / 20 Passed

Description

53%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 concise and third-person with a clear 'what', but it lacks an explicit 'Use when...' trigger clause and omits common trigger terms (refactoring, code review, bug fixes) that the body actually covers. It is mostly distinct thanks to the named tool but has minor overlap with sibling coding-agent skills.

Suggestions

Add an explicit 'Use when...' clause, e.g. 'Use when the user wants to build features, refactor code, review PRs, or batch-fix issues with OpenAI Codex.'

Broaden trigger terms to include natural phrases users say such as 'refactor', 'code review', and 'fix issues', not just 'features, PRs'.

Sharpen the verb beyond the abstract 'Delegate coding' to concrete actions like 'Run OpenAI Codex CLI to build features, refactor, and review PRs.'

Dimension	Reasoning	Score
Specificity	Names the domain ('OpenAI Codex CLI') and 1-2 concrete action hints ('Delegate coding', '(features, PRs)') but the actions are not comprehensive. It is not a 4 because it does not list several specific concrete actions.	3 / 5
Completeness	States a clear 'what' ('Delegate coding to OpenAI Codex CLI') but has no 'Use when...' clause; 'when' is only weakly implied by the '(features, PRs)' parenthetical, which per the rubric caps completeness at 3. Not a 4 because 'when' is not explicit.	3 / 5
Trigger Term Quality	Contains some relevant natural keywords ('coding', 'Codex CLI', 'features', 'PRs') but misses common variations the body covers such as 'refactor', 'code review', and 'bug fixes'. Not a 4 because keyword coverage has clear gaps.	3 / 5
Distinctiveness Conflict Risk	Naming the specific tool 'OpenAI Codex CLI' makes it mostly distinct, but 'delegate coding' overlaps with closely related coding-agent skills (the metadata lists 'claude-code' as related). Not a 5 because of that minor overlap risk.	4 / 5
	Total	13 / 20 Passed

Validation

81%

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 — 13 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning
metadata_field	'metadata' should map string keys to string values	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	13 / 16 Passed

Repository: NousResearch/hermes-agent
Path: skills/autonomous-ai-agents/codex/SKILL.md
Commit: 48f69e5

Reviewed: 4 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.