idea-creator

Generate and rank research ideas given a broad direction. Use when user says "找idea", "brainstorm ideas", "generate research ideas", "what can we work on", or wants to explore a research area for publishable directions.

Quality

74%

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

Fix and improve this skill with Tessl

tessl review fix ./skills/idea-creator/SKILL.md

Quality

Content

62%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 well-engineered, highly actionable workflow with excellent sequencing and validation, but it is verbose and leans on a wide network of shared-references files that are absent from this bundle, leaving the inlined detail heavy and the disclosure shallow.

Suggestions

Consolidate the cross-model/same-family rationale and reviewer-routing/tracing prose into their shared-references files and keep only a one-line pointer in each phase, cutting significant duplication across Phases 1.5, 3, and 4.

Move the citation-discipline, output-protocol, and reviewer-routing detail fully into the referenced files so the body stays an overview rather than inlining their substance.

Verify or ship the referenced shared-references/*.md files so the progressive-disclosure pointers resolve to real content.

Dimension	Reasoning	Score
Conciseness	The ~480-line body is noticeably verbose: it repeatedly restates the cross-model/same-family invariant across Phases 1.5, 3, and 4, duplicates the 'annotate, do not eliminate' guidance for novelty and impact, and re-explains reviewer routing and tracing in multiple sections, padding that Claude could derive from a single statement.	2 / 5
Actionability	It gives mostly executable guidance — concrete MCP call signatures with config JSON, bundle file paths, and a copy-paste-ready IDEA_REPORT.md template — with minor gaps: the landscape-survey and lens fan-out steps describe what to do more than give runnable commands.	4 / 5
Workflow Clarity	The seven-phase pipeline is clearly sequenced with explicit validation checkpoints (objective feasibility gate in Phase 3, cross-model triage before novelty check, pilot timeout kills, Wiki helper WARN-on-missing) and feedback loops for error recovery, matching the top anchor.	5 / 5
Progressive Disclosure	It references many shared-references/.md files (reviewer-routing, fan-out-pattern, acceptance-gate, output-, citation-discipline, integration-contract, review-tracing) but these bundles are not present in the skill directory, so the references are signaled yet unverifiable, and a large amount of routing/citation/tracing detail is inlined rather than split out.	3 / 5
	Total	14 / 20 Passed

Description

86%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 description with excellent trigger coverage and clear what/when completeness, plus a well-scoped niche. Its main weakness is specificity: it surfaces only two actions (generate, rank) while the body performs a richer pipeline (survey, validate, pilot, re-rank).

Suggestions

Add one or two more concrete actions to the description — e.g. 'validate novelty and run cheap pilot experiments' — so the what matches the body's actual pipeline rather than only generate/rank.

Consider naming the output artifact ('writes a ranked idea report to idea-stage/IDEA_REPORT.md') to sharpen the what and reduce overlap with /idea-discovery.

Dimension	Reasoning	Score
Specificity	The description names only two concrete actions — 'Generate and rank research ideas' — for a clear domain, matching the anchor that names the domain plus 1-2 actions but is not comprehensive; it does not list validation, piloting, or ranking-by-evidence actions that the body actually performs.	3 / 5
Completeness	It explicitly answers both 'what' (generate and rank research ideas given a broad direction) and 'when' with concrete trigger phrases, matching the anchor for clear and explicit coverage of both.	5 / 5
Trigger Term Quality	It lists several natural user phrases including synonyms and a non-English term ('找idea', 'brainstorm ideas', 'generate research ideas', 'what can we work on'), giving comprehensive coverage of what users would actually say.	5 / 5
Distinctiveness Conflict Risk	'Research ideas for publishable directions' is a clear niche with distinct triggers and minimal conflict risk, but it sits in a family of related skills (/novelty-check, /research-review, /idea-discovery) creating minor overlap risk, matching the 'mostly distinct; minor overlap' anchor.	4 / 5
	Total	17 / 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
allowed_tools_field	'allowed-tools' contains unusual tool name(s)	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning
relative_links	Relative link issues: 12 suspicious	Warning

	Total	13 / 16 Passed

Repository: wanshuiyin/Auto-claude-code-research-in-sleep
Path: skills/idea-creator/SKILL.md
Commit: f4f20f9

Reviewed: 1 day 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.