CtrlK
BlogDocsLog inGet started
Tessl Logo

memory-merger

Reviews mature LESSONS-LEARNED.md entries, rewrites them as permanent rules in skill/instruction files, archives graduated lessons. Use when graduating lessons into skills, promoting validated lessons, updating skills from past learnings, archiving mature lessons, codifying repeated patterns, or cleaning up a crowded LESSONS-LEARNED.md.

72

Quality

88%

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

SKILL.md
Quality
Evals
Security

Quality

Content

82%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 tight, highly actionable skill body that avoids padding entirely and includes explicit validation gates for a batch operation. The main weaknesses are a subtly broken scan command (rg -c on a single file defeats the awk filter), no error-recovery loop behind the Quality Gates, and a Merge Protocol section that only details workflow steps 3-6.

Suggestions

Fix the citation-count command: 'rg -c' on a single file prints a bare count, so the awk '$2 >= 3' filter never matches. Use 'rg -o "LES-[0-9]+" .opencastle/logs/events.ndjson | sort | uniq -c | sort -rn' (same pattern as the second command) so per-lesson counts are actually produced.

Add an explicit error-recovery loop after the Quality Gates: state what to do when a gate fails (e.g., revert the target-file edit, keep the lesson un-archived, and re-draft), which would lift workflow_clarity to the top anchor.

Add concrete commands for the Quality Gates (e.g., the 'keyword search' gate: 'rg "<rule keyword>" <target file>') and cover workflow steps 1-2 in the Merge Protocol so every numbered step has matching detail.

DimensionReasoningScore

Conciseness

Every token earns its place: a dense criteria table, a six-step numbered workflow, exact edit/attribution/archive templates, and terse anti-patterns — no concept explanations, no padding. Matches anchor 5 ('lean and efficient; assumes Claude's competence'); anchor 4 would require trimmable over-explanation, of which there is none.

5 / 5

Actionability

Mostly copy-paste-ready: exact attribution comment ('<!-- Merged from LES-XXX -->'), exact archive heading format, and concrete rg/awk commands. However, the first scan command is subtly broken — 'rg -c' on a single file prints a bare count, so the 'awk -F: '$2 >= 3'' filter matches nothing and yields no output — and Quality Gates like 'keyword search confirms the merge' give no concrete command. Anchor 4 (concrete commands with minor gaps) fits; not 5 because one provided command is not actually executable as written.

4 / 5

Workflow Clarity

Clear numbered 1-6 sequence plus an explicit Quality Gates checklist (so the destructive/batch cap does not apply), but there is no error-recovery loop — nothing says what to do when a gate fails — and step 6's 'run validation checks' is not concretized into a command. Anchor 4 ('clear sequence with most checkpoints present; minor validation gaps') fits; not 5 because anchor 5 requires feedback loops for error recovery.

4 / 5

Progressive Disclosure

Well-organized single-file skill with clear section headers (Run Criteria, Workflow, Merge Protocol, Quality Gates, Anti-Patterns) and no content that belongs in separate files, though at ~80 body lines it exceeds the under-50-line exception for a top score. The 'Merge Protocol' subsections numbered '### 3' through '### 6' cover only steps 3-6 of the workflow, a minor organizational asymmetry. Anchor 4 (good structure, minor organization gaps) is the best fit.

4 / 5

Total

17

/

20

Passed

Description

95%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: third-person, concrete, and dense with natural trigger phrases and an explicit 'Use when' clause covering six user-intent variants. The only gap is that secondary workflow actions (index updating, validation) are not mentioned, leaving specificity slightly short of comprehensive.

DimensionReasoningScore

Specificity

Lists several specific concrete actions — 'Reviews mature LESSONS-LEARNED.md entries', 'rewrites them as permanent rules in skill/instruction files', 'archives graduated lessons' — but omits parts of the skill's function (index updates, validation gates), so coverage has minor gaps rather than being comprehensive. Not 5 because anchor 5 demands comprehensive coverage of the skill's actions; not 3 because more than 1-2 concrete actions are named.

4 / 5

Completeness

Explicitly answers both questions: the 'what' ('Reviews... rewrites... archives') and a rich explicit 'when' clause ('Use when graduating lessons into skills, ... or cleaning up a crowded LESSONS-LEARNED.md'). Clear what-and-when with concrete trigger phrases is exactly anchor 5; anchor 4 would require a weaker or less specific 'when'.

5 / 5

Trigger Term Quality

Comprehensive natural trigger phrasing including synonyms and the exact file name: 'graduating lessons into skills', 'promoting validated lessons', 'updating skills from past learnings', 'archiving mature lessons', 'codifying repeated patterns', 'cleaning up a crowded LESSONS-LEARNED.md'. Matches anchor 5's standard of natural synonyms plus the file extension/name; no common variation is missing.

5 / 5

Distinctiveness Conflict Risk

Occupies a clear niche — graduation of LESSONS-LEARNED.md entries into permanent skill files — with triggers ('graduating lessons', 'codifying repeated patterns', 'LESSONS-LEARNED.md') unlikely to fire for any other skill. Minimal conflict risk matches anchor 5; anchor 4 would imply overlap with closely related skills, which is not the case here.

5 / 5

Total

19

/

20

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository
monkilabs/opencastle
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.