Selfloop — Development skill for Claude Code
Run a bounded foreground adaptation cycle and stop at evidence-based review.
How to install Selfloop
Installs to ~/.claude/skills/rasputinkaiser-self-improvement-plugin-selfloop/SKILL.md
mkdir -p ~/.claude/skills/rasputinkaiser-self-improvement-plugin-selfloop && curl -fsSL https://raw.githubusercontent.com/RasputinKaiser/Self-Improvement-Plugin/HEAD/commands/selfloop.md -o ~/.claude/skills/rasputinkaiser-self-improvement-plugin-selfloop/SKILL.md Restart Claude Code, or start a new session, for it to be picked up.
What Selfloop does
description: Run a bounded foreground adaptation cycle and stop at evidence-based review.
Follow `references/command-protocol.md`.
Use `goal_state.py selfloop-set ` only for an explicitly requested loop. `status`, `pause`, `resume`, `complete`, and `clear` retain their documented state controls. A request for status never starts a cycle. No background scheduler or automatic child-agent campaign is implied.
For each authorized foreground cycle: inspect `self_correct.py --json` an
Alternatives in Development
- Stop Review Gate — You are reviewing uncommitted changes before the author finishes their turn 664 ★
- Vox Collage Broll — Turn one spoken line into a paper-collage explainer video — Claude Code skill + MaxFusion MCP 198 ★
- Peerbridge MCP — Provider-native adapter and governance layer for auditable multi-agent coding, review, evidence, and release w 136 ★
Full documentation available on GitHub
View Source RepositoryRelated Skills
Scvd Store Execution Contract
A behavioral contract for autonomous work. Name the success criteria before acting, spend the cheapest evidenc
Debug Pro
Resolve existing defects through adaptive evidence gathering, exact reproduction, causal confirmation, bounded
Bake
Implement one milestone (or all sequentially with checkpoints) from the active run's milestones/. Spawns Baker
Flow Ralph
Bounded self-correct loop — fix, verify, refine, max 3 rounds then stop + report
Maintain Check
Run one durable maintainer autopilot cycle — budget-gated, stop-guarded, checkpointed. Use --dry-run to previe
GPU
Diagnose GPU health, discover bounded recipes, and collect typed numeric evidence without guessing from screen
Related Agents
Cook Execute
jeff execute stage for operation tasks. Perform the bounded state transition, preserve the recovery boundary,
Implementer Deep
A higher-effort, foreground implementation role for a genuinely complex, bounded change. Use only instead of —
Hydra Framework Evolution
Evaluate whether a repository-local Hydra change should be promoted into the reusable seed, and record the evi