research-idea

当用户提供任意研究资料、项目背景、实验结果、论文草稿、PR/仓库信息或自然语言线索，希望“提出科学问题”“凝练研究假设”“找创新点”“判断一个想法是否值得做”时使用。该 skill 通过 research-topic-extractor 与 research-literature-review 做查新，并用 parallel-vibe 默认 3 轮串行独立审查打磨多个“科学问题-可证伪假设”候选，最终输出 Research-Idea_{github仓库名}_{pr名}_{时间戳}.md。⚠️ 不适用：用户只需要完整实验方案/分析计划（优先 research-plan）、只要写文献综述正文（优先 research-literature-review）、或只要普通头脑风暴且不需要查新。

Quality

88%

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%

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

The body is a well-sequenced, actionable workflow with strong validation feedback loops and concrete commands/schemas. Its main weaknesses are moderate token redundancy (repeated path templates and install-path variants) and two reference files that are not navigated to from the body while their content is duplicated inline.

Suggestions

Consolidate the recurring .bensz-api/task-…/research-idea/ path template into one defined variable and reference it, and list the three install-location variants once instead of repeating them for init_workspace.py, validate_report.py, and parallel-vibe.

Link references/novelty-check.md and references/agent-review-prompt.md explicitly from steps 3 and 4 ('See references/novelty-check.md for the judgment criteria'), and move the duplicated novelty-tier criteria and review prompt out of the body into those files to avoid inline duplication.

Replace the parallel-vibe loop's '{第 ${round} 轮审查指令}' placeholder with a concrete prompt string or a clear pointer to references/agent-review-prompt.md so the command is closer to copy-paste ready.

Dimension	Reasoning	Score
Conciseness	The body avoids explaining concepts Claude already knows, but the hidden-workspace path template recurs across several sections and the triple system-install path variants (research-idea/scripts/, ~/.codex/skills/, ~/.claude/skills/) repeat for init, validate, and parallel-vibe, adding tokens that could be tightened. Not a 3 because not every token earns its place; not a 1 because there is no concept-fluff padding.	2 / 3
Actionability	Provides fully executable commands (init_workspace.py, the parallel-vibe bash loop with --n and --out-dir, validate_report.py) and concrete JSON schemas with enum fields (theme.json, novelty-decision.json), rather than pseudocode or vague direction.	3 / 3
Workflow Clarity	A clearly sequenced 6-step process with explicit validation checkpoints and feedback loops: init fails early on missing dependencies, fully-researched candidates loop back to step 2, and report validation runs a fix-then-retry cycle before delivery.	3 / 3
Progressive Disclosure	The body is well-sectioned and signals references/report-template.md and the scripts (all real files), but references/novelty-check.md and references/agent-review-prompt.md exist without being linked, and their content is partly duplicated inline. Not a 3 because two reference files are not well-signaled and content that has a dedicated file is repeated in the body.	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 specific, trigger-rich, and self-disambiguating, explicitly stating both what the skill does and when to use it while carving out clear boundaries against neighboring skills. It is a strong, near-exemplary description with no substantive gaps.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — novelty checking via research-topic-extractor/research-literature-review, 3-round parallel-vibe review, and a specifically named Research-Idea_*.md output — well beyond a vague domain mention. Written in third person ('该 skill 通过…'), so no voice penalty applies.	3 / 3
Completeness	Explicitly answers both what ('该 skill 通过…做查新…最终输出…md') and when ('当用户提供…希望…时使用'), with an explicit non-applicability clause; the 'Use when' trigger is present, so it is not capped at 2.	3 / 3
Trigger Term Quality	Covers natural user phrases ('提出科学问题', '凝练研究假设', '找创新点', '判断一个想法是否值得做') plus bilingual metadata keywords, giving good coverage of terms users would actually say.	3 / 3
Distinctiveness Conflict Risk	Clearly delineates the niche from research-topic-extractor, research-literature-review, parallel-vibe, and research-plan, and redirects non-applicable cases, making wrong-skill triggering unlikely.	3 / 3
	Total	12 / 12 Passed

Validation

87%

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 — 14 / 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

	Total	14 / 16 Passed

Repository: huangwb8/ChineseResearchLaTeX
Path: skills/research-idea/SKILL.md
Commit: 01044a6

Reviewed: 3 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.