loop plugins

by OutlineDriven

ODIN workflows for bounded iteration with explicit completion checks.

by pwguler

Skills for Claude Code. Includes 15 skills.

by Mingyang Bao

Iterative TDD review→fix loop — critique, fan out fix agents, verify, gate hard calls to the human, keep tests green, commit per round

by Andrew Vasilyev

Repeats one prompt in fresh claude -p calls until the verification verbs a person names pass, the stated number of iterations has run, the next call could pass the stated budget or two iterations in a row change nothing, with both bounds held outside the model. It keeps nothing in context on every turn.

by Rajdeep Development

Build a new looping AI agent on smoke-monkey-harness (skill + MCP guide server).

by weedmo

Autonomous code improvement loop - autocode: hypothesis-driven parallel experiments with model routing (strategist on the expensive tier, experimenters routed by hypothesis difficulty), a live experiment board via weed-harness loop-report, unlazy-backed termination gates, and kept changes collected as one squash commit each on a worktree branch that ends in a PR.

by weedmo

Autonomous code improvement loop, Claude Code edition - autocode: hypothesis-driven parallel experiments with in-session experimenters routed by the plugin agents, a live experiment board published as an Artifact, unlazy-backed termination gates, and kept changes collected as one squash commit each on a worktree branch that ends in a PR.

by weedmo

Autonomous improvement loops - autocode (code improvement loop with PGE team mode) and auto_research (ML research loop with optional 23-stage pipeline).

by weedmo

Matt Pocock design with the user via matt-design; confirmed-spec autonomous tickets, implementation, verification and independent review via matt-auto. Includes PR babysitting and conflict resolution. Requires weed-harness 6.0+.

by rocky2431

Interview the owner into a grounded goal, then emit the runnable prompt or script that pursues it until an anchor says it is met.

by weedmo

Design artifacts plus shared progress-page delivery, model routing, completion gates, and Claude Code hooks. Design conversations and autonomous execution live in matt-loop.

by Tamas Mezei

A suite for authoring Claude Code artifacts: reviewing-claude-skills and reviewing-claude-subagents each review their artifact in two passes — structure first, stopping at a structural verdict when the shape fails (fit-for-purpose leads the subagent gate), otherwise a full sweep ending in a computed verdict that honors the artifact's recorded waivers; improving-claude-skills and improving-claude-subagents loop the matching review autonomously — review → apply every blocking finding → commit → re-review — until the verdict is acceptable, the findings plateau, or the round cap is hit, asking only one question: the intent brief at kickoff. The subagent loop ends instead of converting when a review says the subagent should be a skill.

by paulkim-lansik

Core loop mechanics: verdict-run (machine-readable PASS/FAIL contract), loop-fix (closed verify->fix loop), lessons (verified-fix memory), classify-risk, require-tests, check-docs-hygiene/check-pr-hygiene/check-module-size (repo hygiene gates). verifier=ceiling — the agent's self-judgement never substitutes for it.

by yukineko

Stuck-loop detector + escalation for Claude Code: a PostToolUse hook that spots repeated identical actions and edit thrash, then injects an escalating nudge to change approach or ask the user for help. Subscription-native (one hook + bundled Rust binary, no API key).