agentsclimarketplace

Complexity audit

Skill yuri-semenenko/ai-engineering-workspace/claude-code/.claude/skills/complexity-audit

One engineering workflow across Claude Code, Codex, Copilot, and Gemini CLI. A portable persona canon, process skills, and safety guardrails, kept in sync by design.

Install
npx -y skills add yuri-semenenko/ai-engineering-workspace --skill complexity-audit

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Scan an entire codebase (or a chosen subtree) for over-engineering — premature abstraction, speculative generality, needless layering, wrapper-only modules, dead config/flags, indirection without payoff — and return a prioritized, deletion-oriented report. Use for "audit complexity", "find over-engineering", "where are we over-built", "complexity-audit". Complements the per-diff /code-review and /simplify (which look at the current change); this looks at the whole tree. Aligned to the persona's anti-pattern list.

SKILL.md

2.5 KB, as published. Nobody here has run it

Complexity Audit

Find accidental complexity that already shipped. This is a standing-codebase scan, not a diff review — use /code-review or /simplify for the current change.

How to run (keep main context lean)

  1. Map the tree first. If a path argument was given, scope the whole audit to that subtree; otherwise take the whole repo. Identify top-level subsystems / packages from the layout (workspaces, src/ domains, services).
  2. Fan out, do not grep inline. Per the persona's session-hygiene rule, dispatch Explore subagents — one per subsystem — each tasked to surface over-engineering only. Each returns a compact findings list, not file dumps. This keeps large intermediate output out of the main conversation.
  3. Dedup and prioritize the merged findings before reporting.

What counts as over-engineering (the persona's anti-patterns)

  • Premature abstraction / speculative generality (interfaces with one implementation, config for things that never vary).
  • Pattern-heavy designs where a function would do (factories, managers, base classes added "for the future").
  • Needless layering / indirection with no payoff; wrapper modules that only re-export.
  • Framework-driven architecture: structure dictated by a framework, not the domain.
  • Dead code, unused flags, config that no longer changes behavior.
  • Tight coupling and leaky abstractions hidden behind "clean" names.

Do not flag: intentional // TRADEOFF(...) annotations (those are tracked by /debt-ledger), validation, error handling, security, or tested invariants.

Output

Prioritized report, most-impactful first. For each finding:

SeverityLocation (file:line)What's over-builtWhy it doesn't pay for itselfProposed simplificationRisk of removing

Severity = blast radius × how speculative it is. End with a short "highest-leverage deletions" shortlist. State assumptions explicitly. No marketing prose.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.