CtrlK
BlogDocsLog inGet started
Tessl Logo

ce-compound-refresh

Refresh the repo's captured learnings against the current codebase. Use when auditing stale, overlapping, superseded, or drifted learnings; avoid general refactor, debugging, or code review unless the learnings store is explicit.

68

Quality

83%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Passed

No findings from the security scan

The canonical home for this skill is ce-compound-refresh in EveryInc/compound-engineering-plugin

SKILL.md
Quality
Evals
Security

Quality

Content

85%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 tightly written orchestration body: lean sections, concrete rules, an unambiguous sequence, and disciplined deferral of detail to ten existing reference files with explicit read triggers. Both sub-5 dimensions are near-misses — actionability defers a few specifics to references by design, and conciseness has minor repeated rationale. Validation and failure handling for this destructive batch workflow are present in the body and its per-action flows.

DimensionReasoningScore

Conciseness

The body is dense and imperative: every section states rules and pointers with no explanation of concepts Claude already knows. Minor trimming is possible — the "store only compounds value" rationale is repeated in the opening and the Discoverability section, and a few justificatory sentences ("A contradiction misleads actively, so it outranks individual staleness") could be shortened. Anchor 4, not 5, because of these small redundancies.

4 / 5

Actionability

Guidance is highly concrete for an instruction-only skill: exact paths (`<root>/solutions/`, `<repo-root>/.compound-engineering/config.yaml`), a copy-paste confirmation prompt with numbered options, a subagent clause to carry verbatim, `git rev-parse --show-toplevel`, and explicit stop conditions. It sits at anchor 4 rather than 5 because some execution specifics ("commit in the repo's convention", validator invocation) are deferred to reference files.

4 / 5

Workflow Clarity

A clearly ordered pipeline (Mode → Worth lens → Artifact Root → Scope → Investigate → Classify → Execute → Vocabulary Capture → Report → Commit → Discoverability Check) with validation checkpoints and error-recovery fallbacks stated in the body: failed writes become Recommended and the run continues, an invalid docs_root stops with a named error, the auto-delete rule carries pre-checks, and the git-failure fallback is signaled. The destructive/batch cap does not apply — validation (bundled validators, pre-checks, confirmation gate) is present.

5 / 5

Progressive Disclosure

Exemplary overview structure: each stage bold-signals "**Read `references/x.md`**" together with what that file covers, and all ten referenced files exist in the bundle. Scripts and assets are reached one level down from the reference files rather than inlined in the body, keeping SKILL.md a lean index with one-level-deep navigation.

5 / 5

Total

18

/

20

Passed

Description

82%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 well-formed description in third-person voice with an explicit Use-when clause, natural trigger synonyms, and an unusually good anti-trigger carve-out that protects against mis-firing on general refactor/debug/review requests. Its only real gap is breadth of action coverage: it summarizes the skill as one composite refresh action rather than naming the concrete actions it performs.

Suggestions

Enumerate the concrete maintenance actions in the description (e.g. "audit each learning, fix stale paths and links, consolidate duplicates, delete superseded docs, and commit the changes") to raise specificity from a single composite action to a multi-action list.

Add one or two common user phrasings as triggers, such as "outdated learnings" or "review the solutions/learnings docs", to broaden trigger coverage beyond stale/overlapping/superseded/drifted.

DimensionReasoningScore

Specificity

"Refresh the repo's captured learnings against the current codebase" names the domain and a single composite action (refresh/audit) rather than a list of distinct concrete actions like fix, consolidate, delete, report, or commit. It matches anchor 3 (domain plus 1-2 concrete actions), not anchor 4, which expects several enumerated actions.

3 / 5

Completeness

It explicitly answers both questions: what ("Refresh the repo's captured learnings against the current codebase") and when ("Use when auditing stale, overlapping, superseded, or drifted learnings"), with concrete trigger phrases and an additional anti-trigger clause. This matches the anchor-5 example pattern.

5 / 5

Trigger Term Quality

"auditing stale, overlapping, superseded, or drifted learnings" supplies four natural synonyms a user would plausibly say. It falls short of anchor 5 because common variations are missing — e.g. "outdated", "review the learnings/solutions docs", or the store's directory name as a trigger.

4 / 5

Distinctiveness Conflict Risk

"avoid general refactor, debugging, or code review unless the learnings store is explicit" actively disambiguates this skill from the broad categories it could be confused with, and the "captured learnings" niche is distinct. Minimal conflict risk.

5 / 5

Total

17

/

20

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.

Validation — 15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

Total

15

/

16

Passed

Repository
crdant/compound-engineering-plugin
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.