CtrlK
BlogDocsLog inGet started
Tessl Logo

meta-optimize

Analyze ARIS usage logs and propose optimizations to SKILL.md files, reviewer prompts, and workflow defaults. Outer-loop harness optimization inspired by Meta-Harness (Lee et al., 2026). Use when user says "优化技能", "meta optimize", "improve skills", "分析使用记录", or wants to optimize ARIS's own harness components based on accumulated experience.

68

Quality

85%

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

73%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, well-sequenced producer skill with strong actionability and validation checkpoints. Its main weakness is verbosity — repeated landing-gate messaging and long boundary/threat-model prose inflate the body without proportional value.

Suggestions

Consolidate the repeated 'never apply in this skill / landing is /meta-apply' message into one canonical statement; Steps 4, 5, and 6 each restate it.

Move the Event Schema Reference block and/or the extended privilege-boundary threat-model prose into a shared-reference file and link to it, shrinking the inline body.

Tighten the model-delta / harness-diet paragraph in Step 1, which restates the 'never deletion candidates' list also given in Key Rules.

DimensionReasoningScore

Conciseness

Mostly efficient and free of basic-concept padding, but the privilege-boundary and threat-model prose is long, and the 'never apply in this skill' landing-gate message is restated several times across Steps 4, 5, and 6, so it could be tightened.

3 / 5

Actionability

Provides executable bash (Step 0 event counting), a copy-paste python json.dumps logger (Step 1.5), a concrete trigger_eval.py command, and a full codex MCP review prompt; held back from 5 by template placeholders ([paste original], [paste diff]) and some procedural-only steps.

4 / 5

Workflow Clarity

A clearly sequenced Step 0→6 pipeline with explicit validation checkpoints (data-sufficiency guard, cross-model advisory review, anti-self-poisoning screen, human landing gate) and a Key Rules checklist; the destructive/batch cap does not bind because verification is pervasive.

5 / 5

Progressive Disclosure

Well-signaled one-level-deep references (shared-references/*.md, tools/*.py, ../meta-apply/SKILL.md) with clear section structure, but no bundle files ship with the skill and some inline material (Event Schema Reference, extended boundary prose) could plausibly live in separate files.

4 / 5

Total

16

/

20

Passed

Description

96%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: concrete actions, explicit bilingual trigger phrases, and a clear what/when structure. The only soft spot is slight overlap risk from the generic term 'improve skills'.

DimensionReasoningScore

Specificity

Lists multiple concrete actions ('Analyze ARIS usage logs and propose optimizations') across several specific targets ('SKILL.md files, reviewer prompts, and workflow defaults'), giving comprehensive coverage rather than a single generic verb.

5 / 5

Completeness

It explicitly answers both 'what' (analyze logs and propose optimizations to named harness components) and 'when' (a concrete 'Use when user says...' clause with specific trigger phrases).

5 / 5

Trigger Term Quality

The 'Use when user says' clause supplies natural trigger phrases in two languages ("优化技能", "meta optimize", "improve skills", "分析使用记录") plus a fallback intent description, covering synonyms a user would actually say.

5 / 5

Distinctiveness Conflict Risk

The ARIS-harness-meta-optimization niche is distinct from regular research workflows, but the broad phrase 'improve skills' carries minor overlap risk with general skill-editing requests, keeping it just below a clean 5.

4 / 5

Total

19

/

20

Passed

Validation

81%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation13 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

allowed_tools_field

'allowed-tools' contains unusual tool name(s)

Warning

frontmatter_unknown_keys

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

Warning

relative_links

Relative link issues: 9 suspicious

Warning

Total

13

/

16

Passed

Repository
wanshuiyin/Auto-claude-code-research-in-sleep
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.