skill-tuning
Universal skill diagnosis and optimization tool. Detect and fix skill execution issues including context explosion, long-tail forgetting, data flow disruption, and agent coordination failures. Supports Gemini CLI for deep analysis. Triggers on "skill tuning", "tune skill", "skill diagnosis", "optimize skill", "skill debug".
76
71%
Does it follow best practices?
Impact
Pending
No eval scenarios have been run
Passed
No known issues
Optimize this skill with Tessl
npx tessl skill review --optimize ./.claude/skills/skill-tuning/SKILL.mdQuality
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 skill description that clearly communicates its purpose, lists specific capabilities with concrete issue types it addresses, and provides explicit trigger terms. It uses proper third-person voice throughout and is concise without being vague. The only minor note is that 'Universal skill diagnosis and optimization tool' is slightly grandiose as an opener, but the rest of the description backs it up with specifics.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: 'Detect and fix skill execution issues including context explosion, long-tail forgetting, data flow disruption, and agent coordination failures.' Also mentions Gemini CLI support for deep analysis. | 3 / 3 |
Completeness | Clearly answers both what ('Detect and fix skill execution issues including context explosion, long-tail forgetting, data flow disruption, and agent coordination failures') and when ('Triggers on "skill tuning", "tune skill", "skill diagnosis", "optimize skill", "skill debug"'). | 3 / 3 |
Trigger Term Quality | Explicitly lists natural trigger terms: 'skill tuning', 'tune skill', 'skill diagnosis', 'optimize skill', 'skill debug'. These are terms users would naturally say when needing this functionality, with good variation coverage. | 3 / 3 |
Distinctiveness Conflict Risk | Occupies a clear niche around skill diagnosis and optimization with very specific trigger terms like 'skill tuning' and 'skill debug' that are unlikely to conflict with other skills. The specific issue types (context explosion, long-tail forgetting) further distinguish it. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
42%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill is well-organized as a navigation hub with excellent progressive disclosure and clear structural overview, but it critically lacks actionable content in the SKILL.md itself. Nearly all executable guidance is deferred to referenced files that aren't provided, making the skill body function more as a table of contents than an operational guide. The workflow is sequenced but missing inline validation details for what are potentially destructive operations.
Suggestions
Add at least one concrete, executable example of a diagnosis action (e.g., actual code or commands for detecting context explosion) directly in SKILL.md rather than deferring everything to referenced files.
Include inline validation/error-recovery steps in the workflow section, e.g., 'If action-verify fails: review state.issues, re-run action-propose-fixes, re-apply' with specific state checks.
Replace the placeholder bash usage examples with a realistic walkthrough showing actual input, diagnosis output, and fix application with expected state.json snippets.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is reasonably structured with tables and diagrams, but includes some redundancy (e.g., the action reference table largely duplicates the workflow table, and the architecture diagram restates what the tables explain). The priority table and key principles are efficient, but overall it could be tightened. | 2 / 3 |
Actionability | Despite extensive structure, the skill provides no executable code, no concrete commands beyond placeholder bash invocations, and no actual implementation of any diagnosis or fix action. Everything actionable is deferred to referenced files (problem-taxonomy.md, tuning-strategies.md, orchestrator.md, action implementations) which are not provided. The SKILL.md itself describes rather than instructs. | 1 / 3 |
Workflow Clarity | The workflow table provides a clear sequence of 8 steps with orchestrator decisions and outputs, and the state management section shows state transitions. However, validation/verification is mentioned only as a step ('action-verify') without explicit feedback loops or error recovery details inline—those are deferred to orchestrator.md. For a process involving destructive changes (applying fixes to skills), the lack of inline validation checkpoints caps this at 2. | 2 / 3 |
Progressive Disclosure | The skill excels at progressive disclosure: it provides a clear overview with well-organized tables and diagrams, then consistently points to one-level-deep references (specs/, phases/, phases/actions/) with a clean reference documents table at the end. Navigation is easy and references are clearly signaled. | 3 / 3 |
Total | 8 / 12 Passed |
Validation
90%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 10 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
allowed_tools_field | 'allowed-tools' contains unusual tool name(s) | Warning |
Total | 10 / 11 Passed | |
- Repository
- catlog22/Claude-Code-Workflow
- Commit
227244f
- 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.