{"slug":"evolve-loop","title":"evolve-loop","description":"A self-learning development pipeline that gets smarter with every cycle. It calibrates itself against your codebase, dynamically selects the highest-impact tasks, implements them in isolated git worktrees, and blocks bad code through eval gates — then learns from each cycle to make better decisions next time.\n\n4 specialized agents (Scout, Builder, Auditor, Operator) work across 5 phases: DISCOVER → BUILD → AUDIT → SHIP → LEARN. The Scout uses multi-armed bandit\nscoring and 8-dimension quality benchmarks to target your codebase's weakest areas. The Builder implements in isolation. The Auditor blocks MEDIUM+ severity\nissues. After each cycle, the pipeline extracts reusable \"instincts\" — patterns about what worked and what failed — that compound over time.\n\nEvery 5 cycles, a meta-cycle evaluates the pipeline itself and evolves its own agent prompts based on measured performance. It ran 151 cycles on its own\ncodebase, shipping 67+ improvements and raising its quality score from 83.8 to 94.4 — all autonomously.\n\nNo external dependencies. MIT licensed.","developer":null,"category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-10-09T12:03:27.111Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/mickeyyaya/evolve-loop.git","web_data":{},"github_data":{},"catalog_sources":{"community":[{"name":"evolve-loop","source":{"sha":"36eebb0512f466510ace20942e046ca7a70ed990","url":"https://github.com/mickeyyaya/evolve-loop.git","source":"url"},"homepage":"https://github.com/mickeyyaya/evolve-loop#readme","description":"A self-learning development pipeline that gets smarter with every cycle. It calibrates itself against your codebase, dynamically selects the highest-impact tasks, implements them in isolated git worktrees, and blocks bad code through eval gates — then learns from each cycle to make better decisions next time.\n\n4 specialized agents (Scout, Builder, Auditor, Operator) work across 5 phases: DISCOVER → BUILD → AUDIT → SHIP → LEARN. The Scout uses multi-armed bandit\nscoring and 8-dimension quality benchmarks to target your codebase's weakest areas. The Builder implements in isolation. The Auditor blocks MEDIUM+ severity\nissues. After each cycle, the pipeline extracts reusable \"instincts\" — patterns about what worked and what failed — that compound over time.\n\nEvery 5 cycles, a meta-cycle evaluates the pipeline itself and evolves its own agent prompts based on measured performance. It ran 151 cycles on its own\ncodebase, shipping 67+ improvements and raising its quality score from 83.8 to 94.4 — all autonomously.\n\nNo external dependencies. MIT licensed."}]}}