compound-refresh

Use when docs/solutions/ learnings may be stale — after refactors, migrations, or dependency upgrades, when a retrieved learning feels outdated or contradicts a recently solved problem, when pattern docs no longer reflect current code, or when reviewing docs/solutions/ for accuracy.

Quality

81%

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.

A thorough, highly actionable maintenance workflow with excellent sequencing, explicit validation gates, and concrete templates. Its weaknesses are repetition of the autonomous-mode and classification rules across many sections, and a monolithic structure that could offload some detail to reference files.

Suggestions

Consolidate the autonomous-mode rules and the Keep/Update/Replace/Archive definitions into single canonical sections, then reference them from Phase 3/Phase 4/Common Mistakes instead of restating them, to reduce token cost.

Move the detailed Phase 3 question-script templates and the full Output Format report skeleton into a reference file (e.g. references/report-format.md) referenced from the main body, to shorten SKILL.md and improve progressive disclosure.

Dimension	Reasoning	Score
Conciseness	The body is purposeful and procedural (not explaining concepts Claude already knows), but the autonomous-mode 'mark as stale when uncertain' rule and the Keep/Update/Replace/Archive model are restated across Mode Detection, Core Rules, Phase 2, Phase 3, Phase 4, and Common Mistakes, so it could be meaningfully tightened.	3 / 5
Actionability	Highly actionable with copy-paste-ready specifics: exact frontmatter fields (status: stale, stale_reason, stale_date: YYYY-MM-DD, archived_date, archive_reason), exact paths (docs/solutions/_archived/), concrete git commands (git branch --show-current), ordered scope-matching strategies, and a full report template covering the common cases.	5 / 5
Workflow Clarity	Clear five-phase sequence (Phase 0–5) with explicit gates and feedback loops: scope routing, evidence-sufficiency assessment before Replace, strict auto-archive criteria (implementation AND problem domain gone), and stale-marking as a fallback loop when evidence is insufficient or writes fail.	5 / 5
Progressive Disclosure	Well-organized single file with clear headers, tables, and easy navigation, and no nested-reference problem; however it is a ~540-line monolith with no external references, and sections like the detailed Phase 3 question templates and report format could arguably live in referenced material, leaving it just below anchor 5.	4 / 5
	Total	17 / 20 Passed

Description

81%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 trigger-rich, highly distinctive description that excels at telling Claude when to invoke it, with strong natural phrasing and low conflict risk. Its main weakness is that the concrete 'what does this do' actions are implied through trigger clauses rather than stated up front.

Suggestions

Lead the description with a short action clause (e.g., 'Reviews docs/solutions/ learnings against the current codebase and refreshes stale pattern docs.') before the 'Use when...' triggers, so the 'what' is as explicit as the 'when'.

Add one or two more concrete actions (e.g., 'marks ambiguous learnings as stale', 'archives obsolete docs') to lift specificity toward comprehensive coverage.

Dimension	Reasoning	Score
Specificity	Names the domain (docs/solutions/ learnings, pattern docs) and a couple of concrete actions (reviewing for accuracy, refreshing pattern docs), but the description is dominated by trigger conditions rather than a comprehensive list of concrete actions, so it stops at anchor 3.	3 / 5
Completeness	The 'when' is exceptionally explicit and concrete, and the 'what' is implied via 'reviewing docs/solutions/ for accuracy' and 'pattern docs no longer reflect current code', but the 'what' is not front-loaded as a clear action statement, so it falls just below anchor 5.	4 / 5
Trigger Term Quality	Comprehensive natural trigger phrases a user would actually say — 'learnings may be stale', 'after refactors, migrations, or dependency upgrades', 'retrieved learning feels outdated or contradicts a recently solved problem', 'pattern docs no longer reflect current code' — with synonyms (stale/outdated) covering the space.	5 / 5
Distinctiveness Conflict Risk	A clear niche (maintaining docs/solutions/ learnings and derived pattern docs) with distinct, specialized triggers and minimal overlap risk with other skills.	5 / 5
	Total	17 / 20 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (550 lines); consider splitting into references/ and linking	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: lucianghinda/superpowers-ruby
Path: skills/compound-refresh/SKILL.md
Commit: 99fac3d

Reviewed: 6 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.