try-works/role-model / .agents/skills/recursive-mode
recursive-mode
UnverifiedRepository workflow orchestration skill for staged implementation, locked artifacts, late-phase receipts, and durable memory maintenance. Use when executing recursive-mode runs, resuming a run, locking a phase, or verifying locks.
.recursive/AGENTS.md
View on GitHub# AGENTS.md
<!-- RECURSIVE-MODE-AGENTS:START -->
## .recursive AGENTS Router
This file is a lightweight routing/index doc for agents already working inside the repository.
It exists to reduce blind doc-by-doc scanning. It is not a second workflow spec.
## Canonical Rule
- Treat `/.recursive/RECURSIVE.md` as the single workflow source of truth.
- If this file conflicts with `/.recursive/RECURSIVE.md`, follow `/.recursive/RECURSIVE.md`.
## Suggested Read Order
1. Read `/.recursive/RECURSIVE.md` first for workflow rules and required behavior.
2. Read `/.recursive/STATE.md` when the current repo state matters.
3. Read `/.recursive/DECISIONS.md` when prior rationale or relevant earlier work matters.
4. Read `/.recursive/memory/MEMORY.md` when task context may depend on durable memory.
5. Read `/.recursive/memory/skills/SKILLS.md` when the task may use delegated review, subagents, review bundles, smoke-harness portability work, or other capability-sensitive execution.
6. Read the recursive-mode package README or maintainer notes from the installed skill directory or source package checkout when changing the package itself.
## Task Routing
- Starting or resuming a recursive-mode run:
- `/.recursive/RECURSIVE.md`
- `/.recursive/STATE.md`
- `/.recursive/DECISIONS.md`
- `/.recursive/memory/MEMORY.md`
- Authoring a new recursive-mode spec or `00-requirements.md`:
- `/.recursive/STATE.md`
- `/.recursive/DECISIONS.md`
- `/.recursive/memory/MEMORY.md`
- the installed `recursive-spec` skill
- relevant code and tests for the requested area
- Benchmarking recursive-mode against a non-recursive baseline:
- Install the separate optional `recursive-benchmark` add-on only when the user explicitly asks for benchmarking.
- Prefer `find-skills` when available; otherwise use `npx skills add <recursive-benchmark-package-or-repo> --full-depth`.
- The default exported `recursive-mode` package intentionally excludes benchmark fixtures and benchmark skill files.
- After the benchmark add-on is installed, follow its packaged fixture and harness docs.
- Working on reusable package/bootstrap/docs for this repo:
- the recursive-mode package README or maintainer notes from the installed skill directory or source package checkout
- the recursive-mode installer scripts from the installed skill directory or source package checkout
- Working on phase artifact structure or lint expectations:
- the recursive-mode artifact template from the installed skill directory or source package checkout
- the recursive-mode lint/status helpers from the installed skill directory or source package checkout
- Working on delegated review, subagent behavior, or routed CLI delegation:
- `/.recursive/memory/skills/SKILLS.md`
- `/.recursive/config/recursive-router.json`
- `/.recursive/config/recursive-router-discovered.json`
- the installed `recursive-router`, `recursive-subagent`, and `recursive-review-bundle` skills
- Working on memory behavior:
- `/.recursive/memory/MEMORY.md`
- the installed `recursive-training` skill
- `/.recursive/scripts/recursive-training-loader.py`
- `/.recursive/memory/training/`
- `/.recursive/memory/skills/SKILLS.md`
## Non-Canonical Bridges
These are adapters, not second specs:
- `/.codex/AGENTS.md`
- `/AGENTS.md`
- `/.agent/PLANS.md`
Read them only when the tool or host expects those entrypoints.
<!-- RECURSIVE-MODE-AGENTS:END -->
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |