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.
70
85%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Passed
No known issues
Quality
Discovery
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.
This is a strong description that clearly articulates what the skill does (chains two specific sub-skills into a pipeline), what the output is (focused proposal + experiment roadmap), and when to use it (with multiple natural trigger phrases including multilingual terms). It is distinctive, complete, and specific with minimal room for improvement.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | The description lists concrete actions: chains 'research-refine' and 'experiment-plan', transforms a vague research direction into a focused final proposal plus detailed experiment roadmap. These are specific, concrete outcomes. | 3 / 3 |
Completeness | Clearly answers both 'what' (chains research-refine and experiment-plan into an end-to-end workflow producing a focused proposal and experiment roadmap) and 'when' (explicit 'Use when' clause with multiple trigger scenarios). | 3 / 3 |
Trigger Term Quality | Includes strong natural trigger terms: 'one-shot pipeline', 'end-to-end', 'build a pipeline', 'generate both the method and experiment plan together', and even the Chinese term '串起来'. Good coverage of how users would naturally phrase this request. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive — it specifically describes a chained pipeline of two named sub-skills ('research-refine' and 'experiment-plan'), clearly differentiating it from either individual skill. The 'end-to-end' and 'pipeline' framing further reduces conflict risk. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
70%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a well-structured orchestration skill with excellent workflow clarity and progressive disclosure. Its main weakness is moderate actionability—it describes what to do at each phase but delegates the concrete execution details to sibling skills without providing inline executable examples or commands. Conciseness could be improved by eliminating redundancy between Phase 1 exit criteria and Phase 2 gate check.
Suggestions
Consolidate the Phase 1 exit criteria and Phase 2 gate check into a single checklist to eliminate redundancy and improve conciseness.
Add a concrete example of how to invoke or trigger the `research-refine` and `experiment-plan` sub-workflows (e.g., slash commands, function calls, or explicit instructions) to improve actionability.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is reasonably efficient but includes some redundancy—e.g., the Phase 2 gate check repeats items already listed as Phase 1 exit criteria. Some sections like 'Key Rules' contain guidance Claude could infer (e.g., 'keep the main paper story compact'). Overall mostly lean but could be tightened. | 2 / 3 |
Actionability | The workflow provides clear sequencing and specific output file paths, and the PIPELINE_SUMMARY.md template is concrete and copy-paste ready. However, the actual method refinement and experiment planning steps delegate to sibling skills without providing executable commands or concrete examples of how to invoke them. The guidance is structured but largely descriptive rather than executable. | 2 / 3 |
Workflow Clarity | The five-phase workflow is clearly sequenced with explicit gate checks (Phase 2 Planning Gate), conditional logic (Phase 0 triage for reuse vs re-run), and clear exit criteria for Phase 1. The 'If in doubt, prefer re-running' heuristic and the 'If the verdict is still REVISE' conditional provide good error recovery guidance. Validation is embedded at the gate check step. | 3 / 3 |
Progressive Disclosure | The skill clearly positions itself as an orchestration layer, with well-signaled one-level-deep references to sibling skills (`../research-refine/SKILL.md`, `../experiment-plan/SKILL.md`) and shared protocols. The instruction 'read these sibling skills only when needed' is excellent progressive disclosure. Content is appropriately split between overview and detail. | 3 / 3 |
Total | 10 / 12 Passed |
Validation
100%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
- Commit
a425a71
- 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.