Almost-autonomous agent harness that takes a GitHub issue or Linear story through planning, review, implementation, hardcore code review, and draft-PR creation. Self-tunes its subagents via a meta-reflection loop within hard safety boundaries.
★not rated 2 3mo agoA
tokens not measured
originalMIT
Autonomous loop that fetches unresolved PR review threads, fixes them via pr-comment-fixer, hardens the diff via hardcore-code-reviewer, runs the host repo's resolved verification sequence, pushes, resolves threads, re-requests copilot review, and repeats until the PR is review-clean or hard caps trip.
★not rated 2 3mo agoA
tokens not measured
originalMIT
Evaluates findings and ideas from conversations and creates high-value GitHub issues. Checks for duplicates before creating, and updates existing issues with new findings instead of creating redundant ones.
★not rated 2 3mo agoA
tokens not measured
originalMIT
Strict hardcore code reviewer that spawns 13 parallel subagents to review code from every angle: bugs, security, architecture, tests, error handling, performance, complexity, observability, API contracts, data/migrations, accessibility, type safety, and doc/comment drift.
★not rated 2 3mo agoA
tokens not measured
originalMIT
A persistent, indexed knowledge bank that stores learnings, ADR decisions, technical context, and session notes as markdown files inside your repo. Scales to thousands of files with fast grep-based index lookups. Integrates with Claude hooks for automatic session capture.
★not rated 2 3mo agoA
tokens not measured
originalMIT
Spawns 11 parallel agents to audit a codebase for technical debt, groups findings into actionable issues, and offers GitHub issue creation.
★not rated 2 3mo agoA
tokens not measured
originalMIT
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: