vibe-check
Judgment linter for vibe-coded output — reads the energy of the code, not just correctness. Use when the user says "vibe check", "check this vibe code", "does this hold up", "sanity check this AI code", or after a fast generation session before committing.
76
70%
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 ./skills/vibe-check/SKILL.mdQuality
Discovery
54%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
The description excels at trigger terms with natural, user-facing phrases that would help Claude select this skill at the right time. However, it critically lacks specificity about what the skill actually does — 'reads the energy of the code' is buzzword-heavy fluff that doesn't describe concrete actions like identifying patterns, flagging issues, or producing specific outputs. The vague capability description undermines both its usefulness and its distinctiveness from general code review skills.
Suggestions
Replace 'reads the energy of the code, not just correctness' with specific concrete actions, e.g., 'Flags common AI-generated code smells such as hallucinated APIs, redundant logic, missing error handling, and inconsistent patterns'.
Clarify what output or analysis the skill produces, e.g., 'Produces a structured review with severity ratings and suggested fixes'.
Differentiate more clearly from general code review by specifying the unique aspects of vibe-coded/AI-generated code analysis.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | The description uses vague, abstract language like 'reads the energy of the code' and 'judgment linter' without listing any concrete actions. It doesn't specify what the skill actually does — no mention of specific checks, outputs, or analysis steps. | 1 / 3 |
Completeness | The 'when' is explicitly and thoroughly covered with a 'Use when...' clause containing multiple triggers. However, the 'what' is extremely vague — 'reads the energy of the code, not just correctness' doesn't explain what the skill concretely does or produces. | 2 / 3 |
Trigger Term Quality | Includes multiple natural trigger phrases users would actually say: 'vibe check', 'check this vibe code', 'does this hold up', 'sanity check this AI code', and the situational trigger 'after a fast generation session before committing'. These are realistic and varied. | 3 / 3 |
Distinctiveness Conflict Risk | The 'vibe check' and 'vibe code' triggers are fairly distinctive, but 'sanity check this AI code' and 'does this hold up' could easily overlap with general code review or linting skills. The vague capability description makes it harder to distinguish from other code quality tools. | 2 / 3 |
Total | 8 / 12 Passed |
Implementation
85%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a strong, well-structured skill that provides clear, actionable guidance for a judgment-based code review pass. The assertion categories table is excellent — specific enough to guide behavior while requiring genuine reasoning. Minor weaknesses include some unnecessary framing in the overview and language-specific footgun lists that Claude already knows, but overall the skill earns its token budget.
Suggestions
Trim the overview paragraph — the metaphorical framing ('what's the energy like') and positioning against static linters is context Claude doesn't need. Start directly with what vibe-check does.
Remove or significantly condense the language-specific examples in 'Read the Room' (Python mutable defaults, JS == vs ===, etc.) — these are well-known to Claude and contradict the skill's own guideline that findings should require judgment beyond what linters catch.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is mostly efficient and well-structured, but includes some unnecessary framing ('a way of asking what's the energy like right now?') and the overview paragraph explains the concept more than needed. The language-specific examples in 'Read the Room' are things Claude already knows. However, the tables and structured output format are lean. | 2 / 3 |
Actionability | The skill provides a concrete workflow with specific steps, clear output format with examples showing exact file:line reference patterns, specific assertion categories with precise definitions, and concrete severity defaults. The example findings are copy-paste ready templates that show exactly what output should look like. | 3 / 3 |
Workflow Clarity | The 5-step workflow is clearly sequenced (Gather → Read → Report → Summarize → Escalate), with an explicit escalation threshold (8+ findings or 4+ categories). The process is well-defined with clear entry/exit conditions and the summary line serves as a validation checkpoint. For a single-pass judgment tool, this is appropriately thorough. | 3 / 3 |
Progressive Disclosure | The skill is well-organized with clear sections (Overview, Input, Categories, Severity, Workflow, Guidelines, See Also). The 'See Also' section provides one-level-deep references to related skills with clear descriptions of when to use each. Content is appropriately contained within a single file given its scope. | 3 / 3 |
Total | 11 / 12 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 — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
- Repository
- tslateman/duet
- Commit
c3b1fc2
- 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.