deep-research

多源深度调研管道（Web Deep Research + Coder 合成 + 云端模型咨询）。 Use when: 技术问题需要多源调查、设计决策需要证据、operator说"调研"/"research"、需要咨询云端模型。 Not for: 简单搜索（直接用 WebSearch）、已有结论的确认。 Output: 调研报告 + 证据合成或咨询文档（含回填区）。

Quality

82%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

High

Do not use without reviewing

Quality

Content

65%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.

The body is highly actionable with concrete selectors, templates, and path conventions, but is verbose for its token budget and ships references to files that are not present in the bundle. Workflow sequencing is solid though some validation checkpoints are implicit.

Suggestions

Consolidate the duplicate 常见错误 tables and move time-stamped verification notes (2026-03-10 实测) into a single 'deprecated/old' section to recover conciseness.

Add the missing bundle files (refs/chatgpt-browser-automation.md, refs/claude-ai-browser-automation.md, refs/gemini-browser-automation.md, cat-cafe-skills/refs/research-prompt-template.md) or remove the references so progressive disclosure is backed by real files.

Add explicit failure-recovery checkpoints to the browser-automation wait/poll steps (e.g., what to do if the stop button never disappears) to lift workflow clarity above 2.

Dimension	Reasoning	Score
Conciseness	Mostly actionable project-specific procedure rather than generic concept explanation, but the ~315-line bilingual body carries redundancy (two 常见错误 sections, repeated Quota/verification notes) and time-stamped checks ('2026-03-10 实测验证') that bloat tokens without being in a deprecated section.	2 / 3
Actionability	Concrete DOM selectors, JS file-injection snippets, exact path conventions, and a fill-in Part 1 prompt template make the browser-automation and consultation workflows copy-paste ready.	3 / 3
Workflow Clarity	Step 1–5 (Mode A) and Step 1–3 (Mode B) are clearly sequenced with validation (Step 4 codebase check, Step 5 commit-SHA verification), but browser-automation waiting/polling steps lack failure-recovery checkpoints and some validation is only implicit.	2 / 3
Progressive Disclosure	References are clearly signaled and one level deep ('详见 refs/chatgpt-browser-automation.md'), but the referenced files (refs/*.md, cat-cafe-skills/refs/research-prompt-template.md) do not exist in any bundle directory, so the split is not actually backed by bundle structure and inline detail that could live in refs remains in SKILL.md.	2 / 3
	Total	9 / 12 Passed

Description

100%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.

The description is specific, complete, and distinctive, with natural trigger terms and an explicit use/not-for/output structure. It cleanly satisfies each dimension at the top anchor.

Dimension	Reasoning	Score
Specificity	Names the domain and multiple concrete capabilities — 'Web Deep Research + Coder 合成 + 云端模型咨询' and an explicit 'Output: 调研报告 + 证据合成或咨询文档（含回填区）'.	3 / 3
Completeness	Explicitly answers both 'what' (多源深度调研管道) and 'when' via the 'Use when:' clause, plus a 'Not for:' boundary and expected output.	3 / 3
Trigger Term Quality	Trigger list ('调研', 'research', '深度研究', '问一下 GPT Pro', '咨询云端') covers natural phrasings users would actually say, including common Chinese variations.	3 / 3
Distinctiveness Conflict Risk	A clear niche (triangulated multi-source deep research + cloud-model consultation) with a 'Not for:' clause distinguishing it from plain WebSearch and Explore agents, unlikely to mis-trigger.	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
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: zts212653/clowder-ai
Path: cat-cafe-skills/deep-research/SKILL.md
Commit: 80782c5

Reviewed: 11 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.