Agent skill
elegance-pipeline
Run a persistent multi-agent code-elegance workflow with 4 scouts, 2 judges, 1 planner, 1 verifier, and 1 gated implementer. Use when evaluating source-file elegance across a repository and optionally converting weaknesses into refactor work.
Install this agent skill to your Project
npx add-skill https://github.com/ANcpLua/ancplua-claude-plugins/tree/main/plugins/elegance-pipeline/skills/elegance-pipeline
SKILL.md
Elegance Pipeline
Multi-agent code-elegance workflow with persistent state and stage gates.
Pipeline stages
4 Scouts (parallel, sonnet) -> 2 Judges (parallel, opus) -> 1 Planner (opus)
-> 1 Verifier (opus) -> 1 Implementer (gated, opus)
State manager
All orchestration goes through the pipeline state manager:
python ${CLAUDE_PLUGIN_ROOT}/elegance_pipeline/pipeline.py [--state-dir <dir>] <command>
Commands: init, status, prompt, submit, signal
Default state lives at .claude/elegance_pipeline/state/ and is shared by default.
If you want one dedicated team per spec, give each run its own --state-dir.
How to orchestrate
- Run
statusto see which slots are ready - For each ready slot, run
prompt --role <role> --slot <slot>to get the task prompt - Spawn the appropriate fully qualified subagent (
elegance-pipeline:elegance-scout,elegance-pipeline:elegance-judge,elegance-pipeline:elegance-planner,elegance-pipeline:elegance-verifier, orelegance-pipeline:elegance-implementer). If your runtime exposes a different fully qualified name, use that exact identifier instead of the bare short name. - After each agent completes, submit its output via the state manager
- Run
statusagain to see what unlocked - Repeat until the pipeline is complete or blocked
Orchestration rules
- Scouts run in parallel (all 4 at once via background agents)
- Judges run in parallel after ALL scouts are submitted
- Planner runs after BOTH judges are submitted
- Verifier runs after the planner is submitted
- Implementer only runs when the implementation signal is READY
- Never bypass stage gates
- If the verifier says no implementation is warranted, stop
Agent model mapping
| Role | Agent | Model | Access |
|---|---|---|---|
| Scout | elegance-pipeline:elegance-scout | sonnet | read-only |
| Judge | elegance-pipeline:elegance-judge | opus | read-only |
| Planner | elegance-pipeline:elegance-planner | opus | read-only |
| Verifier | elegance-pipeline:elegance-verifier | opus | read-only |
| Implementer | elegance-pipeline:elegance-implementer | opus | full edit |
Manual signal override
python ${CLAUDE_PLUGIN_ROOT}/elegance_pipeline/pipeline.py [--state-dir <dir>] signal on
python ${CLAUDE_PLUGIN_ROOT}/elegance_pipeline/pipeline.py [--state-dir <dir>] signal off
If Connectors Available
githubPost elegance score deltas as PR review commentslinearCreate refactor tasks from planner output when implementer gate opensslackNotify on pipeline completion with verdict and scout findings summaryjiraLog verifier decision and refactor scope as a Jira ticket
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
example-skill
Brief description of what this skill does. Use when [specific trigger situations].
carlini-jr
Leaderless multi-agent swarm with Playwright oracle. Spawns N workers in isolated worktrees with pre-partitioned DOD items. Workers implement and verify via screenshots. No orchestrator, no shared state.
otel-expert
Use this when working with OpenTelemetry, telemetry, observability, traces, spans, metrics, logs, OTLP, semantic conventions, or instrumentation. Triggers on questions like "what attributes should I use", "how do I configure the collector", "what's the semconv for X".
design-studio
Design intelligence studio. Creative direction + data-driven recommendations for distinctive frontend interfaces. 50 styles, 97 palettes, 57 font pairings, 99 UX guidelines, 25 chart types, 13 stacks (React, Next.js, Vue, Svelte, SwiftUI, React Native, Flutter, Tailwind, shadcn/ui, Astro, Nuxt, Jetpack Compose). Actions: plan, build, create, design, implement, review, fix, improve, optimize, enhance, refactor, check UI/UX code. Projects: website, landing page, dashboard, admin panel, e-commerce, SaaS, portfolio, blog, mobile app, .html, .tsx, .vue, .svelte. Styles: glassmorphism, claymorphism, minimalism, brutalism, neumorphism, bento grid, dark mode, responsive, skeuomorphism, flat design. Topics: color palette, accessibility, animation, layout, typography, font pairing, spacing, hover, shadow, gradient.
writing-rules
Guide for creating hookify rules to prevent unwanted AI behaviors. Use when user asks to create, write, or configure hookify rules.
hades
IF cleanup/elimination needed THEN use this. IF zero suppressions THEN this. IF dead code THEN this. IF duplication THEN this. IF frontend design quality audit THEN use this with --goggles. Smart-Hades: every session gets a Smart ID, deletion permit, and audit ledger. Team lead skill — spawns 4 debate teammates per phase (+3 goggles teammates when equipped). Ignores public API, semver, changelog. Pure functional destruction. Idempotent: same input, same output. Requires CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1.
Didn't find tool you were looking for?