By wilrf
Post-refactoring codebase cleanup: detects stale/dead code via git history and dependency analysis, simplifies active code, and archives unused files with a full paper trail.
Simplifies and refines active code for clarity, consistency, and maintainability while preserving exact functionality. Focuses only on files classified as FRESH by the staleness-analyzer. Reads project standards from CLAUDE.md. Use after the staleness analysis to clean up the code that is staying.
Archives files classified as DEAD by the staleness-analyzer. Moves files to an _archive/ directory preserving their original path structure, updates all references, and maintains an ARCHIVE_LOG.md with full provenance. Use after the staleness-analyzer has produced its inventory and the user has approved the archive candidates.
Analyzes the codebase to classify files into three categories: fresh (recently modified and actively used), stale (not modified recently but still referenced), and dead (unreferenced, untouched, safe to archive). Uses git log, import/export graphs, and framework entrypoint awareness. Produces a scored inventory. Use proactively after any major refactoring session.
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
AI agents perform better when they have personalities. Paranoid agents find bugs others miss. Obsessive agents find bloat others ignore.
Paranoid hunters find bugs others miss.
Calm fixers apply minimal, surgical repairs.
Obsessive optimizers find every wasted byte and millisecond.
Bug Hunting:
/bughunt → BUGHUNT.md → /bugfix → BUGFIX.md + fixed code
Optimization:
/speedrun → baseline metrics → hunters → SPEEDRUN.md → fixes → verification
/plugin marketplace add wilrf/wilrf-marketplace
/plugin install bughunt@wilrf-marketplace
/plugin install bugfix@wilrf-marketplace
/plugin install speedrun@wilrf-marketplace
| Hunter | Archetype | Focus |
|---|---|---|
| security | PARANOID | XSS, injection, secrets |
| auth | DISTRUSTFUL | Sessions, tokens, perms |
| edge-case | OBSESSIVE | Null, empty, boundaries |
| error-handling | PESSIMISTIC | Try/catch, boundaries |
| test | CYNICAL | Flaky, coverage, assertions |
| env | CAUTIOUS | Config, env vars |
| frontend | EMPATHETIC | React, CSS, a11y |
| performance | IMPATIENT | N+1, memory, bundle |
| backend | SKEPTICAL | API, logic, data |
| database | SUSPICIOUS | Queries, RLS, N+1 |
| type-safety | PEDANTIC | TypeScript, any |
| api | METICULOUS | Endpoints, validation |
| dependency | WARY | Outdated, vulnerabilities |
| Fixer | Archetype | Approach |
|---|---|---|
| security | CALM | Minimal secure patches |
| auth | TRUSTING-BUT-VERIFYING | Verify then fix |
| edge-case | PRAGMATIC | Simple defensive code |
| error-handling | OPTIMISTIC | Graceful handling |
| test | CONSTRUCTIVE | Improve tests minimally |
| env | CONFIDENT | Simplify config |
| frontend | OBJECTIVE | Minimal UI fixes |
| performance | PATIENT | Measured optimizations |
| backend | TRUSTING | Minimal logic fixes |
| database | METHODICAL | Careful schema/query fixes |
| type-safety | PRACTICAL | Minimal type fixes |
| api | EFFICIENT | Minimal endpoint fixes |
| dependency | DECISIVE | Clean dependency changes |
| Hunter | Archetype | Focus |
|---|---|---|
| bundle | OBSESSIVE | File sizes, tree-shaking, code splitting |
| complexity | PEDANTIC | Cyclomatic complexity, nesting, length |
| dead-code | SUSPICIOUS | Unused exports, orphan files |
| dependency | WARY | Bloat, duplicates, vulnerabilities |
| build | IMPATIENT | Compilation time, caching |
| algorithm | ANALYTICAL | Big O, data structures |
| query | SKEPTICAL | N+1, indexes, slow queries |
| web-vitals | EMPATHETIC | LCP, CLS, INP, TTFB |
| memory | PARANOID | Leaks, unbounded caches |
| image | METICULOUS | Formats, compression, lazy loading |
Find every bug. Miss nothing. Three passes minimum.
The best diff: +100 -1000
The Iron Law: NO OPTIMIZATION CLAIM WITHOUT MEASUREMENT
| Plugin | Output | Purpose |
|---|---|---|
| bughunt | BUGHUNT.md | Bugs with severity, evidence, confidence |
| bugfix | BUGFIX.md | Fixes applied, false positives rejected |
| speedrun | SPEEDRUN.md | Optimizations with ROI, before/after metrics |
npx claudepluginhub wilrf/wilrf-marketplace --plugin codebase-simplifySurgical bug fixing with validation-first approach. Reads BUGHUNT.md and applies minimal, tactical fixes. Prefers deletion over addition.
Psychologically-enhanced bug hunting with personality-primed parallel agents. Each hunter has a distinct archetype and cognitive style for deeper, more thorough bug detection.
Parallel optimization agents that find performance bottlenecks, reduce bundle sizes, eliminate dead code, and verify improvements with before/after metrics.
The autonomous build engine (build mode of the agent-loops toolkit): given any task, designs a gate that can prove it done (and can fail), gets human sign-off on the target, then builds and self-verifies against that gate until green — stopping at a committed slice. Install the companion agent-loops (foundation playbook), review, planning, and infra plugins for the other modes.
Find and remove dead code across the codebase
Point it at a codebase and it figures out what's worth doing — bugs, perf, tech debt, what to build next — then writes plans any agent can execute. It can hand the work to a cheaper model and review the result. It never edits your code.
Finds real bottlenecks. DB N+1, memory leaks, blocking calls, re-renders.
コードベースの不要なフォールバック・互換性コード・デッドコードを 3つの専門エージェント(scanner + historian + verifier)が協調して検出・安全性検証する監査スキル。 scanner が自律探索で候補を発見し、historian が git 履歴から時間軸の文脈を付与し、
A collection of Claude Code skills — autonomous optimization, multi-goal orchestration, and more
Tactical code cleanup: type strictness, single responsibility, and fail-fast patterns. Detects sloppy code, workarounds, and silent failures.