research-refine-pipeline

Run an end-to-end workflow that chains `research-refine` and `experiment-plan`. Use when the user wants a one-shot pipeline from vague research direction to focused final proposal plus detailed experiment roadmap, or asks to "串起来", build a pipeline, do it end-to-end, or generate both the method and experiment plan together.

Quality

82%

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

81%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 orchestration skill: a clear phased workflow with real validation gates, concrete output paths, and a ready-to-fill summary template. Its main weaknesses are minor redundancy at the edges and total dependence on external sibling/shared files whose paths are relative and absent from this bundle.

Suggestions

Trim the overlap between the "Composing with Other Skills" section, the final Key Rules bullet, and the description's scope statement — one canonical disambiguation is enough.

Verify the referenced paths exist relative to the deployed layout (../research-refine/SKILL.md, ../experiment-plan/SKILL.md, ../../shared-references/output-*.md), or inline a minimal fallback for the three output protocols so the skill is self-contained.

Deduplicate the Phase 4 template's Contribution Snapshot and Must-Prove Claims sections against the Phase 1 exit criteria and Phase 2 gate answers, referencing them instead of restating.

Dimension	Reasoning	Score
Conciseness	Mostly lean with every template and checklist earning its place, but there is minor redundancy: the "Composing with Other Skills" section repeats the scope split already stated in the description and the last Key Rules bullet, and the Phase 4 template's Contribution Snapshot / Must-Prove Claims sections recap Phase 1-2 exit criteria.	4 / 5
Actionability	Concrete guidance throughout — explicit output paths, per-phase exit criteria, a copy-paste-ready PIPELINE_SUMMARY.md template, and a verbatim user-facing summary — but the actual mechanics of both sub-workflows are fully delegated to sibling skill files that are not part of this bundle, leaving a dependency gap.	4 / 5
Workflow Clarity	Phases 0-5 are clearly sequenced with explicit validation checkpoints: Phase 0 stale/reuse triage of FINAL_PROPOSAL.md, Phase 1 explicit exit criteria, and Phase 2's gate check with "If these answers are not crisp, tighten the final proposal first" — a genuine validate-then-retry loop before the experiment stage.	5 / 5
Progressive Disclosure	Stage detail is correctly deferred with well-signaled, one-level-deep references ("read these sibling skills only when needed", purpose-labeled output-protocol links), but the referenced paths (../research-refine/SKILL.md, ../experiment-plan/SKILL.md, ../../shared-references/*.md) are not bundled with this skill, making the navigation unverifiable or broken in a standalone copy.	4 / 5
	Total	17 / 20 Passed

Description

83%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: it clearly states what the skill does (chains two named workflows into a one-shot pipeline) and gives an explicit, natural-language Use-when clause with varied triggers including a Chinese synonym. The only weaknesses are minor — a slightly thin deliverable enumeration and small overlap risk with the staged sibling skills.

Dimension	Reasoning	Score
Specificity	Names two concrete sub-workflows it chains (`research-refine` and `experiment-plan`) and concrete deliverables ("focused final proposal plus detailed experiment roadmap"), matching the anchor for several specific actions with minor gaps — the full deliverable list lives in the body rather than the description.	4 / 5
Completeness	Explicitly answers both what ("Run an end-to-end workflow that chains research-refine and experiment-plan… final proposal plus detailed experiment roadmap") and when ("Use when the user wants a one-shot pipeline… or asks to…") with concrete trigger phrases, matching the top anchor exactly.	5 / 5
Trigger Term Quality	Includes natural triggers users would say — "asks to '串起来'", "build a pipeline", "do it end-to-end", "generate both the method and experiment plan together" — with good coverage including a non-English synonym, though a few variants (e.g. "one-shot", "full pipeline" as triggers) are absent.	4 / 5
Distinctiveness Conflict Risk	The integrated-pipeline niche is distinct and the when-clause ("one-shot pipeline… both… together") steers away from the staged sibling skills, but there is minor overlap risk with `research-refine` and `experiment-plan` themselves since it reuses their trigger vocabulary.	4 / 5
	Total	17 / 20 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
relative_links	Relative link issues: 3 suspicious	Warning

	Total	15 / 16 Passed

Repository: wanshuiyin/Auto-claude-code-research-in-sleep
Path: skills/skills-codex/research-refine-pipeline/SKILL.md
Commit: 2132036

Reviewed: 4 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.