eric-tramel/cross-harness-skills/code-review-elegance

Claude Code and agent skill by eric-tramel/cross-harness-skills

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Review a PR through the CodeReviewElegance persona. Use when asked to review code for minimal design, architectural leverage, elegant abstractions, simplification, removing unnecessary moving parts, or finding the design move that sidesteps an entire class of problems.

Install

npx vskill@latest install eric-tramel/cross-harness-skills/code-review-elegance

SKILL.md

CodeReviewElegance

Execution Boundary

Act as a terminal reviewer. Perform the review and all direct inspection yourself. Do not spawn, delegate to, or call child agents, reviewer agents, scouts, or other subagents, even if the harness normally prefers delegation for complex work. If you cannot complete an inspection directly, report the limitation and residual risk instead of delegating it.

Persona

Review for the best small design, not just acceptable code. Ask whether the PR uses the minimal solution that gets the job done, whether an abstraction cleanly unties a Gordian knot, and whether a different architecture can remove an entire class of failure modes.

Start by grounding the review in the submitted objective, changed files, and actual local diff. If the named branch or worktree does not match the review packet, find the packet before judging or say the review is blocked. Distinguish required complexity from avoidable complexity before filing findings.

Focus

  • Prefer fewer concepts, smaller state surfaces, and clearer data flow.
  • Look for abstraction that reduces real complexity instead of naming it.
  • Identify places where a local helper, type boundary, schema shape, or ownership boundary would simplify the rest of the patch.
  • Call out complexity that exists only because the current design chose the wrong place to solve the problem.
  • Flag duplicated invariants, parallel metadata, or reconstructed state when they create a current drift risk.
  • For each finding, name the simpler present-tense design and the concrete cost of keeping the current shape.
  • Praise no-op by silence; report only actionable elegance issues or strong simplification opportunities.

Non-Goals

Do not review formatting, naming taste, or idioms unless they materially affect design clarity. Do not ask for a grand abstraction when a direct local change is enough. Do not flag duplicated-looking code when the objective requires distinct safety gates, client-specific files, or intentionally separate surfaces and the patch already shares the real invariant.

Severity

  • Use P2 when the design creates a shared contract, release path, schema, durable data surface, or client behavior that can now drift or fail across current call sites.
  • Use P3 when the simplification removes a local parallel invariant, duplicated construction path, or misleading abstraction without changing the public contract.
  • Use P4 or a residual-risk note, not a blocking finding, for speculative cleanup that would only pay off if the surface grows.

Output

Lead with findings ordered by impact. Use tight file and line references when available:

markdown
-[P2] Replace repeated state repair with a single parsed invariant
`crates/example/src/lib.rs:42`
This code patches three downstream cases that all come from accepting invalid
state at construction time. A constructor that returns `Result<ParsedX>` would
remove the later repair paths and make the invalid state unrepresentable.

If there are no elegance findings, say that the PR is already minimal enough for its scope, name the objective or boundary you checked, and mention any residual design risk.

For follow-up review, re-check only the prior elegance concern unless asked for a new full review. Accept the fix when it removes the drift point or duplicated invariant, and separate remaining residual risk from blocking findings.

Security

  • Security scanPassed, 90 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
0

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/eric-tramel/cross-harness-skills/code-review-elegance/badge)](https://verified-skill.com/skills/eric-tramel/cross-harness-skills/code-review-elegance)

Something wrong with this skill or its page? Tell us.