deep-research

Run a disciplined, multi-source research investigation for a high-stakes question or decision — fan-out web search across many channels, parallel sub-agents, source triangulation (each claim backed by ≥3 independent sources), an adversarial review pass, and every source saved to its own file with verbatim quotes for reuse. Use when a low-quality answer is expensive: strategy work, comparing N products/methods/markets, validating a hypothesis with external data, or mapping how a field works. NOT for quick fact-checks (answer directly), structured 12-dimension competitor scoring (use competitive-teardown), or fast topic overviews where the decision risk is low (use the research router instead).

Quality

87%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

—

The risk profile of this skill

Quality

Content

75%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.

The body is a well-structured, auditable methodology with a clear 9-phase pipeline, explicit validation checkpoints, and a sensible split between the in-skill operating discipline and the upstream catalog. It is slightly redundant between its comparison and mechanisms sections and could spell out error-recovery loops more crisply.

Suggestions

Consolidate the feature list so triangulation, parallel sub-agents, adversarial pass, and refresh appear once — fold 'How it differs' into 'Core mechanisms' to remove the duplicated enumeration.

Make the error-recovery loop explicit at the Verify and Score phases (e.g., 'if a thesis lacks ≥3 sources, loop back to Phase 4 with refined queries before proceeding') to reach anchor-5 workflow clarity.

Add one concrete sub-agent dispatch example or prompt template in Phase 4 so the parallel fan-out is copy-paste actionable rather than only described.

Dimension	Reasoning	Score
Conciseness	The body is lean and assumes Claude's competence — no padding about what research or sub-agents are — but the 'How it differs from a quick research router' and 'Core mechanisms' sections restate the same features (triangulation, parallel sub-agents, adversarial pass, refresh) twice, which could be tightened.	4 / 5
Actionability	Concrete, specific guidance throughout — named files (plan.md, sources/NN.md, refresh_targets.md), explicit thresholds ('≥3 independent, differently-typed sources'), and scoring axes (Credibility/Recency/Bias) — but no sample sub-agent dispatch prompts or exact commands, leaving minor execution gaps for an instruction-only skill.	4 / 5
Workflow Clarity	The 9-phase pipeline table is clearly sequenced with depth gating and real checkpoints (Phase 5 triangulation flagging, Phase 6.5 'Verify — Lightweight citation check before closing', 're-evaluate between rounds'), but the error-recovery loop is softer than the anchor-5 ideal of explicit 'if errors: fix and re-validate'.	4 / 5
Progressive Disclosure	Well-organized sections with a clearly signaled one-level reference (references/full-catalog.md) that defers the fast-moving catalog upstream rather than inlining it; minor gap is that the reference file then points further to several upstream files, a light second hop (external).	4 / 5
	Total	16 / 20 Passed

Description

100%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.

The description is exemplary: it states concrete capabilities, gives natural trigger phrases for both inclusion and exclusion, and explicitly disambiguates from neighboring research skills. Third-person voice is maintained throughout.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'fan-out web search across many channels, parallel sub-agents, source triangulation', 'an adversarial review pass', and 'every source saved to its own file with verbatim quotes' — giving comprehensive coverage rather than a generic verb.	5 / 5
Completeness	Explicitly answers both what ('Run a disciplined, multi-source research investigation...') and when ('Use when a low-quality answer is expensive: strategy work, comparing N products/methods/markets, validating a hypothesis...'), with concrete trigger phrases and exclusion guidance.	5 / 5
Trigger Term Quality	Natural trigger phrases a user would actually say are well covered with synonyms: 'strategy work', 'comparing N products/methods/markets', 'validating a hypothesis', 'mapping how a field works', plus explicit negation triggers ('quick fact-checks', 'competitor scoring', 'fast topic overviews').	5 / 5
Distinctiveness Conflict Risk	Clear niche with distinct triggers and minimal conflict risk — it explicitly differentiates from sibling skills ('NOT for... use competitive-teardown', 'use the research router instead'), making misrouting unlikely.	5 / 5
	Total	20 / 20 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: alirezarezvani/claude-skills
Path: research/deep-research/skills/deep-research/SKILL.md
Commit: 19392f7

Reviewed: ~24 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.