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.
docs/templates/commands/recursive-review-bundle.md
View on GitHub# recursive-review-bundle (prompt template)
> Paste this into chat if your agent does not support custom slash commands.
## Usage Pattern
```text
Generate a recursive review bundle for run <run-id> phase 03.5 using the canonical bundle format.
```
## Script (Recommended)
```powershell
python "<SKILL_DIR>/scripts/recursive-review-bundle.py" --repo-root . --run-id "<run-id>" --phase "03.5 Code Review" --role code-reviewer --artifact-path "/.recursive/run/<run-id>/03.5-code-review.md" --upstream-artifact "/.recursive/run/<run-id>/00-requirements.md" --upstream-artifact "/.recursive/run/<run-id>/02-to-be-plan.md" --upstream-artifact "/.recursive/run/<run-id>/03-implementation-summary.md" --routing-config-path ".recursive/config/recursive-router.json" --routing-discovery-path ".recursive/config/recursive-router-discovered.json" --routed-cli "codex" --routed-model "gpt-5.4" --audit-question "Which R# remain incomplete?" --audit-question "Which changed files drift from the plan?" --required-output "Findings ordered by severity"
pwsh -NoProfile -File "<SKILL_DIR>/scripts/recursive-review-bundle.ps1" -RepoRoot . -RunId "<run-id>" -Phase "03.5 Code Review" -Role code-reviewer -ArtifactPath "/.recursive/run/<run-id>/03.5-code-review.md" -UpstreamArtifact "/.recursive/run/<run-id>/00-requirements.md","/.recursive/run/<run-id>/02-to-be-plan.md","/.recursive/run/<run-id>/03-implementation-summary.md" -PriorRef "/.recursive/memory/MEMORY.md" -RoutingConfigPath ".recursive/config/recursive-router.json" -RoutingDiscoveryPath ".recursive/config/recursive-router-discovered.json" -RoutedCli "codex" -RoutedModel "gpt-5.4" -AuditQuestion "Which R# remain incomplete?","Which changed files drift from the plan?" -RequiredOutput "Findings ordered by severity"
```
Example for a Phase 4 test-review bundle:
```powershell
python "<SKILL_DIR>/scripts/recursive-review-bundle.py" --repo-root . --run-id "<run-id>" --phase "04 Test Summary" --role tester --artifact-path "/.recursive/run/<run-id>/04-test-summary.md" --upstream-artifact "/.recursive/run/<run-id>/02-to-be-plan.md" --upstream-artifact "/.recursive/run/<run-id>/03-implementation-summary.md" --evidence-ref "/.recursive/run/<run-id>/evidence/logs/green/tests.log" --audit-question "Are the executed tests sufficient for the changed behavior?" --required-output "Audit recommendation"
```
## What It Writes
- A markdown review bundle under `/.recursive/run/<run-id>/evidence/review-bundles/`
- Artifact content hash for stale-bundle detection
- Diff basis from `00-worktree.md`
- Current changed file list
- Upstream artifact list
- Relevant addenda auto-discovered by default
- Prior recursive evidence refs when supplied
- Code refs, evidence refs, audit questions, and required output hints
- Routed review metadata such as `Routing Config Path`, `Routing Discovery Path`, `Routed CLI`, and `Routed Model`
## Phase 3.5 Expectation
When review is delegated, record the generated bundle path in `03.5-code-review.md`:
```md
## Review Metadata
- Review Bundle Path: `/.recursive/run/<run-id>/evidence/review-bundles/<bundle>.md`
- Routing Config Path: `/.recursive/config/recursive-router.json`
- Routing Discovery Path: `/.recursive/config/recursive-router-discovered.json`
- Routed CLI: `codex`
- Routed Model: `gpt-5.4`
```
If repairs materially change scope, changed files, or evidence, regenerate the bundle before re-audit.
If meaningful subagent work follows from the delegated review, capture that work in a matching action record under `/.recursive/run/<run-id>/subagents/`.
The maintainer smoke harness uses the same bundle contract, so this file is also the reference shape for automated regression runs.
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 |