improve-skill

Turn plugin-eval findings into a concrete rewrite brief for a Codex skill. Use when the user already evaluated a skill and now wants Codex to improve it, especially after asking what to fix first.

Quality

83%

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

77%Weight 40%Scale 1-3

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

A compact, well-sequenced instruction skill with a genuine verify-then-compare feedback loop, but it is held back by a non-portable absolute path in its workflow and a broken reference to a missing bundle file.

Suggestions

Replace the hardcoded '/Users/benlesh/.codex/skills/skill-creator/SKILL.md' path in step 3 with a portable reference (e.g., a relative path or a documented lookup) so the step is executable in any environment.

Either create the referenced '../../references/chat-first-workflows.md' file (and the references/ directory) or remove/fix the dead link so progressive disclosure points to real resources.

Expand the actionability of step 2 with a concrete cue for how to split required vs. recommended fixes (e.g., a brief criterion), since it currently relies on judgment alone.

Dimension	Reasoning	Score
Conciseness	The body is lean and free of padding or explanations of concepts Claude already knows, with terse bullets and direct trigger phrasing where every token earns its place.	3 / 3
Actionability	Steps 1 and 4 give concrete, copy-paste-ready commands ('plugin-eval analyze <skill-path> --brief-out <brief.json>'), but step 3 references a hardcoded user-local absolute path ('/Users/benlesh/.codex/skills/skill-creator/SKILL.md') that is non-portable and effectively non-executable outside that machine.	2 / 3
Workflow Clarity	The four-step workflow is clearly sequenced and includes an explicit validation/feedback loop ('Re-run the evaluation and compare before and after outputs'), meeting the score-3 bar with a concrete verification checkpoint.	3 / 3
Progressive Disclosure	The single reference ('../../references/chat-first-workflows.md') is structurally one level deep, but no references/, scripts/, or assets/ directories exist and the path does not resolve to a real file, so the navigation points to a missing resource.	2 / 3
	Total	10 / 12 Passed

Description

90%Weight 40%Scale 1-3

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

A strong, well-scoped description with an explicit trigger clause and natural user phrasings. Its only weakness is specificity: it foregrounds one main action rather than enumerating several concrete capabilities.

Dimension	Reasoning	Score
Specificity	It names the domain and a concrete primary action ('Turn plugin-eval findings into a concrete rewrite brief for a Codex skill') but centers on a single action rather than listing multiple distinct concrete operations like the score-3 anchor.	2 / 3
Completeness	An explicit 'Use when...' clause answers both what it does (turn findings into a rewrite brief) and when to invoke it (after evaluation, when wanting improvement or asking what to fix first), which is the score-3 pattern.	3 / 3
Trigger Term Quality	Natural user phrasings appear ('wants Codex to improve it', 'asking what to fix first'), matching real things a user would say, and the body lists recognizable chat requests like 'Improve this skill' and 'What should I fix first'.	3 / 3
Distinctiveness Conflict Risk	The niche is tightly scoped to the plugin-eval / Codex-skill improvement workflow via distinctive terms ('plugin-eval findings', 'rewrite brief'), making it unlikely to fire for unrelated skills.	3 / 3
	Total	11 / 12 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: openai/plugins
Path: plugins/plugin-eval/skills/improve-skill/SKILL.md
Commit: 11c74d6

Reviewed: 17 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.