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

88%

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

77%

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 clear multi-step workflows and feedback loops, but it is overlong with conversational padding and a FastClaw tangent, and it carries dead references to bundle files (agents/*.md, eval-viewer/generate_review.py) that do not exist.

Suggestions

Trim or move the FastClaw runtime section and conversational asides ("Cool? Cool.", the plumbers/grandparents paragraph, repeated core-loop restatement) to bring the body under the 500-line guideline.

Resolve dead references: either add the missing agents/grader.md, agents/comparator.md, agents/analyzer.md and eval-viewer/generate_review.py to the bundle, or point to the scripts/ equivalents (e.g., scripts/generate_report.py) and scripts/utils.py that actually exist.

Split the large inline "Running and evaluating test cases" and "Description Optimization" sections into references/ files, leaving SKILL.md as an overview with one-level-deep pointers.

Dimension	Reasoning	Score
Conciseness	The ~527-line body exceeds the skill's own 500-line guideline and includes unnecessary padding (the 40-line FastClaw detour, the plumbers/grandparents demographic paragraph, "Cool? Cool.", "Good luck!", and a repeated restatement of the core loop), though much of the procedural core is genuinely valuable and not derivable.	2 / 3
Actionability	Concrete executable commands ("python -m scripts.aggregate_benchmark <workspace>/iteration-N"), complete JSON templates, exact directory layouts, and precise field-name requirements ("text", "passed", "evidence") make the guidance copy-paste ready.	3 / 3
Workflow Clarity	The eval workflow is broken into an explicit Step 1-5 sequence with a feedback loop (run -> grade -> review feedback -> improve -> rerun into iteration-N+1), explicit sequencing warnings ("don't stop partway through", "only opportunity to capture this data"), and clear stopping criteria.	3 / 3
Progressive Disclosure	A "Reference files" section clearly signals references and most paths resolve (references/schemas.md, assets/eval_review.html, scripts/*.py), but agents/grader.md, agents/comparator.md, agents/analyzer.md and eval-viewer/generate_review.py are referenced yet absent from the bundle, and the monolithic over-500-line body inlines content that could be split out.	2 / 3
	Total	10 / 12 Passed

Description

100%

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 well-crafted: third-person voice, explicit both-what-and-when structure, multiple concrete actions, and natural trigger terms covering the full skill lifecycle. No first/second-person voice or vague fluff.

Dimension	Reasoning	Score
Specificity	"Create new skills, modify and improve existing skills, and measure skill performance" lists multiple concrete actions (create, modify, improve, measure) rather than vague language, matching the score-3 anchor.	3 / 3
Completeness	It explicitly answers what ("Create new skills, modify and improve... and measure skill performance") and when ("Use when users want to create a skill from scratch, edit, or optimize..."), with explicit trigger guidance.	3 / 3
Trigger Term Quality	"create a skill from scratch, edit, or optimize an existing skill, run evals, benchmark skill performance" are natural phrases a user would actually say, with good coverage including common variations.	3 / 3
Distinctiveness Conflict Risk	The skill-creation/improvement/eval niche has distinct triggers unlikely to overlap with other skills; "run evals to test a skill" and "optimize a skill's description" are highly specific to this domain.	3 / 3
	Total	12 / 12 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 (528 lines); consider splitting into references/ and linking	Warning

	Total	15 / 16 Passed

Repository: fastclaw-ai/fastclaw
Path: internal/agent/bundled_skills/skill-creator/SKILL.md
Commit: 533a138

Reviewed: about 18 hours 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.