CtrlK
BlogDocsLog inGet started
Tessl Logo

deep-research

Multi-agent research orchestration: split a research goal into parallel sub-goals, run each via headless `claude -p` subprocesses, aggregate results into a polished report file. Use for systematic web/document research, competitive or industry analysis, batch link/dataset processing, and long-form evidence synthesis. Triggers: "深度调研", "deep research", "wide research", "多 Agent 调研", "系统调研".

72

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Low

Low-risk findings worth noting

SKILL.md
Quality
Evals
Security

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 thoroughly sequenced, highly actionable orchestration workflow with strong validation checkpoints, but it is monolithic with no progressive disclosure and carries some repetition plus generic guidance that inflates token use.

Suggestions

Move the 'Claude Code 非交互模式参考' appendix and the generic '思考与写作指南' into a separate reference file (e.g. references/claude-noninteractive.md) and link to it, so SKILL.md stays a lean overview.

State the search-service priority (skills → firecrawl → exa → WebFetch/WebSearch) once as a single canonical rule and reference it elsewhere instead of restating it in four places.

Extract the dispatch/parallel bash templates into a scripts/ file and reference them, reducing inline duplication between the process steps and the examples section.

DimensionReasoningScore

Conciseness

Mostly efficient with concrete value, but the search-service priority rule ('skills → firecrawl → exa → WebFetch/WebSearch') is restated 4+ times and the generic '思考与写作指南' plus the largely-known 'Claude Code 非交互模式参考' appendix add tokens that could be tightened.

2 / 3

Actionability

Provides concrete, copy-paste-ready bash templates with real flags (`claude -p`, `--allowedTools`, `--output-format json`, `timeout`, `tee`, GNU Parallel), concrete directory conventions, and worked scheduling examples — fully executable rather than pseudocode.

3 / 3

Workflow Clarity

A strictly numbered 0–9 end-to-end sequence with explicit validation checkpoints (摸底 evidence gate, user-confirmation gate, pre-dispatch prompt review, double-QA before delivery) and feedback loops for failures/timeouts, plus a final self-check checklist.

3 / 3

Progressive Disclosure

No bundle files exist (references/, scripts/, assets/ absent) and all content — including the long reference appendix and best-practices section — is inline in one monolithic file; content that should be split out (e.g., the Claude Code non-interactive reference) is not separated or signaled.

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 strong: concrete capabilities, explicit 'Use for...' triggers, and a distinct, well-defined niche. It fully satisfies the what/when/trigger requirements with no obvious gaps.

DimensionReasoningScore

Specificity

Lists multiple concrete actions — 'split a research goal into parallel sub-goals, run each via headless `claude -p` subprocesses, aggregate results into a polished report file' — matching the 'Lists multiple specific concrete actions' anchor rather than the partial level-2 anchor.

3 / 3

Completeness

Clearly answers both 'what' (split / run subprocesses / aggregate) and 'when' ('Use for systematic web/document research, competitive or industry analysis, batch link/dataset processing, and long-form evidence synthesis') with explicit triggers, satisfying the level-3 anchor.

3 / 3

Trigger Term Quality

Explicit Triggers ('深度调研', 'deep research', 'wide research', '多 Agent 调研', '系统调研') are natural phrasings a user would actually say, giving good coverage of common variations rather than only jargon or generic terms.

3 / 3

Distinctiveness Conflict Risk

The niche — multi-agent research orchestration via headless `claude -p` subprocesses — is clearly distinct with dedicated triggers, making it unlikely to fire for the wrong skill.

3 / 3

Total

12

/

12

Passed

Validation

93%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

allowed_tools_field

'allowed-tools' contains unusual tool name(s)

Warning

Total

15

/

16

Passed

Repository
feiskyer/claude-code-settings
Reviewed

Table of Contents

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.