review
Framework for code review that captures context future maintainers need—concerns raised, alternatives rejected, risks accepted. Use for PRs, local changes, or architecture review when the decision matters more than the diff. Produces structured feedback with must-address issues, suggestions, and observations "for the record."
77
71%
Does it follow best practices?
Impact
Pending
No eval scenarios have been run
Advisory
Suggest reviewing before use
Optimize this skill with Tessl
npx tessl skill review --optimize ./skills/review/SKILL.mdQuality
Discovery
85%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 description that clearly articulates a distinctive niche—decision-context-preserving code review rather than generic code review. It effectively communicates both what the skill does and when to use it, with specific output categories. The main weakness is moderate trigger term coverage, missing some common user phrasings like 'pull request' or 'review my code'.
Suggestions
Add more natural trigger term variations such as 'pull request', 'review my code', 'merge request', or 'code changes' to improve discoverability when users use common phrasings.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: captures concerns raised, alternatives rejected, risks accepted; produces structured feedback with must-address issues, suggestions, and observations. These are concrete, actionable outputs. | 3 / 3 |
Completeness | Clearly answers both what (captures context for future maintainers, produces structured feedback with categorized issues) and when ('Use for PRs, local changes, or architecture review when the decision matters more than the diff'). The 'Use for...' clause explicitly defines trigger scenarios. | 3 / 3 |
Trigger Term Quality | Includes some natural keywords like 'code review', 'PRs', 'architecture review', and 'diff', but misses common variations users might say such as 'pull request', 'review my code', 'code feedback', or 'merge request'. The phrase 'decision matters more than the diff' is evocative but not a natural trigger term. | 2 / 3 |
Distinctiveness Conflict Risk | The focus on decision-context preservation and structured feedback categories (must-address, suggestions, observations 'for the record') creates a clear niche distinct from generic code review or linting skills. The emphasis on capturing rationale rather than just finding bugs is highly distinctive. | 3 / 3 |
Total | 11 / 12 Passed |
Implementation
57%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a solid instructional skill that provides a clear review framework with a useful output template and good cross-references. Its main weaknesses are moderate verbosity (explaining concepts Claude already understands like what 'correctness' and 'maintainability' mean in code review) and some sections that describe rather than instruct. The workflow would benefit from explicit validation checkpoints to ensure review completeness.
Suggestions
Trim the evaluation dimensions (Correctness, Design, Maintainability, Risk) to just the non-obvious checklist items — Claude already knows what these concepts mean; focus on project-specific or easily-missed aspects.
Add a validation checkpoint before finalizing the review, e.g., 'Before posting: verify every Must Address item includes a concrete suggestion or alternative, and the Concerns for the Record section is non-empty.'
Remove or significantly condense the 'Output Quality' section — it restates principles already embedded in the template and workflow.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is reasonably well-structured but includes some unnecessary explanatory content that Claude already knows (e.g., explaining what correctness, design, maintainability mean, the Bacchelli & Bird research citation, and the 'Output Quality' section which restates obvious review principles). The anti-patterns section, while useful, borders on teaching Claude things it already understands. | 2 / 3 |
Actionability | The review template is concrete and copy-paste ready, and the GitHub CLI commands are executable. However, much of the guidance remains at the level of checklists and principles rather than specific executable steps. The 'Local Code Review' and 'Architecture Review' sections are vague lists rather than concrete procedures. | 2 / 3 |
Workflow Clarity | The 4-step workflow is clearly sequenced and the review template provides good structure. However, there are no validation checkpoints or feedback loops — no step says 'verify your review covers all dimensions before posting' or 'if context is insufficient, loop back to step 1.' For a process that produces structured output, there's no verification that the output is complete. | 2 / 3 |
Progressive Disclosure | The content is well-organized with clear sections, appropriate use of headers, and well-signaled cross-references to related skills (/naming, /adr, /prose, FRAMEWORKS.md, RECIPE.md). Content is appropriately scoped for a single SKILL.md file without being monolithic. | 3 / 3 |
Total | 9 / 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.