skill-creator

Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.

Quality

81%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Quality

Content

70%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.

A thorough, well-sequenced skill-creation workflow with strong validation feedback loops and mostly executable guidance. Its weaknesses are length/conciseness (over its own 500-line budget with repeated material) and broken or missing bundle references that undermine the progressive-disclosure structure.

Suggestions

Trim the body below 500 lines by removing the duplicate core-loop restatement at the end (lines 602-613) and condensing the "Communicating with the user" and "How skill triggering works" sections, or move that detail into a reference file.

Ship the missing referenced files (`eval-viewer/generate_review.py` and the `agents/grader.md`, `agents/comparator.md`, `agents/analyzer.md` docs) or remove/replace the references so navigation is not broken.

Add explicit "read this when..." guidance next to each reference link (as done for schemas.md) so the one-level-deep pointers are clearly signaled.

Dimension	Reasoning	Score
Conciseness	The body runs ~613 lines, exceeding the 500-line limit the skill itself prescribes, and contains some unnecessary or repeated material — the core loop is restated at the top (lines 11-29) and again verbatim near the end (lines 602-613), and sections like "Communicating with the user" and "How skill triggering works" explain concepts Claude largely already knows.	3 / 5
Actionability	Most guidance is executable — concrete `host.skills.*` API calls, exact `python -m scripts.aggregate_benchmark`/`scripts.run_loop` invocations with flags, and JSON templates — but several commands retain `<placeholder>` values and a few referenced executables (e.g. `eval-viewer/generate_review.py`) are not confirmed present, leaving minor gaps from fully copy-paste-ready.	4 / 5
Workflow Clarity	The process is laid out as a clear, well-sequenced loop (Capture Intent → Interview → Write SKILL.md → Test Cases → Step 1-5 run/evaluate → Improve → iterate) with explicit validation/feedback checkpoints (grade → aggregate → analyst pass → viewer review → user feedback → revise) and gating on publish/packaging steps.	5 / 5
Progressive Disclosure	Structure is reasonable and uses one-level-deep references to `references/schemas.md`, `assets/eval_review.html`, and `scripts/*`, but several referenced files are missing from the bundle (`eval-viewer/generate_review.py`, `agents/grader.md`, `agents/comparator.md`, `agents/analyzer.md`) and the body exceeds its own 500-line guideline with content that could be split into reference files.	3 / 5
	Total	15 / 20 Passed

Description

92%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.

A strong, third-person description that covers both what the skill does and when to use it with concrete trigger phrases and multiple specific actions. The only minor gap is trigger-term coverage, which misses a few natural synonyms and casual phrasings.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete, distinct actions — "Create new skills, modify and improve existing skills, and measure skill performance" plus "edit, or optimize", "run evals", "benchmark ... with variance analysis", and "optimize a skill's description" — giving comprehensive coverage of the skill's capabilities.	5 / 5
Completeness	It clearly answers what the skill does (create/modify/improve/measure) and explicitly answers when Claude should use it via a concrete "Use when users want to..." clause enumerating trigger scenarios.	5 / 5
Trigger Term Quality	The "Use when" clause includes good natural phrasings users would say ("create a skill from scratch", "edit", "optimize an existing skill", "run evals", "benchmark"), but a few common variants (e.g. "make a skill", "skill builder") are missing and there are no file-extension-style triggers, so it sits just below fully comprehensive.	4 / 5
Distinctiveness Conflict Risk	It carves a clear niche — skill authoring and performance measurement — with distinct, meta-level triggers ("create a skill", "run evals", "benchmark skill performance") that are unlikely to collide with other skills.	5 / 5
	Total	19 / 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
skill_md_line_count	SKILL.md is long (614 lines); consider splitting into references/ and linking	Warning

	Total	15 / 16 Passed

Repository: UnicomAI/wanwu
Path: configs/microservice/bff-service/configs/agent-skills/claude-science/skill-creator/SKILL.md
Commit: 99df6e4

Reviewed: 9 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.