菜单
Use when starting a new evolve-loop cycle and the cycle goal is ambiguous or only described as a one-line objective. Generates a structured specification before any code is written. Surfaces assumptions explicitly.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Use when the user invokes /evo:loop or asks to run autonomous improvement cycles, self-evolving development, compound discovery, or multi-cycle code improvement with research, build, audit, and learning phases
Use after build has produced build-report.md. Validates the build via four parallel sub-auditors (eval-replay, lint, regression, build-quality) and produces ALL-PASS verdict. Adversarial mode default-on per CLAUDE.md.
Use when the user asks to explain how a feature/subsystem works and how the loops built it, or what the concurrent loops/lanes/cycles did and why. Default output = per-FEATURE HTML explainer pages (two movements — how it works today, then how it got built, with verified file:line claims and drift callouts) under docs/explain/; per-batch/per-lane story pages are the fallback framing. Derived from geoffreylitt/explain-diff-html, customized for evolve-loop.
Use when writing or changing any production code or tests — new features, bug fixes, refactors, or test authoring, in any language, by any model on any CLI.
Use when running any LLM agent below Fable 5 — Claude (Opus/Sonnet/Haiku), GPT via codex, Gemini via agy, or local models via ollama — to adopt Fable 5's operating discipline — evidence-first investigation, premise verification, root-cause-only fixes, adversarial self-review, calibrated autonomy, and honest failure reporting. Load at session start or as a persona overlay for phase agents on any CLI.
Use after ship completes. Three sub-reflectors (instinct, gene, failure) run in parallel to extract lessons, update gene pool, and analyze any failures. Off the latency-critical path.
| name | scout |
| description | Use when starting a new evolve-loop cycle and the cycle goal is ambiguous or only described as a one-line objective. Generates a structured specification before any code is written. Surfaces assumptions explicitly. |
Sprint 3 composable skill (v8.16+). Inspired by
addyosmani/agent-skills/spec-driven-development. Invoked as/evo:scoutand called by theloopmacro.
failedApproaches retry queue (the spec already exists)| Step | Action | Exit criteria |
|---|---|---|
| 1 | Read cycle goal from .evolve/state.json | Goal extracted |
| 2 | List explicit assumptions about scope, stack, constraints | At least 3 numbered assumptions |
| 3 | Identify dependency graph (what depends on what) | Dependency tree drawn |
| 4 | Surface unknowns + risks | Risks enumerated with mitigation hints |
| 5 | Write spec to <workspace>/spec.md | File present, fresh, contains challenge token |
Spec at <workspace>/spec.md with sections:
## Goal (one paragraph)## Assumptions (numbered list)## Dependency Graph (text or ASCII)## Risks (table: risk / likelihood / mitigation)## Out of Scope (explicit deferrals)| Fact | Value |
|---|---|
| Phase | scout (plan archetype, mandatory) |
| Persona | agents/evolve-scout.md |
| Profile | .evolve/profiles/scout.json — CLI claude-tmux, tier balanced, fan-out ×3 |
| Inputs | intent.md |
| Artifact | scout-report.md (cycle workspace) |
scout-report.md must declare:
## Selected Tasks (also accepted: ## Proposed Tasks)Invoked by:
/evo:scout (when scout-codebase sub-scout runs)loop macro skill at calibrate→research transitionThis skill does not invoke other personas. It is a workflow used by Scout-class personas.
For deeper detail see skills/loop/phase1-research.md. The spec format follows addyosmani/agent-skills/spec-driven-development conventions.