Run it inside Claude Code, then install plugins from it with /plugin.
Plugins
1best-of-nRun a Best of N tournament. Single pass: N parallel solutions judged blind by an Opus reviewer who ranks them and picks a winner. Two pass: the reviewer also distills lessons, a second round of N guided attempts runs, and a final Opus ranker picks the winner from the new attempts plus the carried-over round-one winner
2farnsworth-loopRun a Farnsworth Loop tournament via the explicit marker 'farnsworth loop:N' (single pass) or 'farnsworth loop:N:2' (two pass). Same engine as best-of-n: N parallel attempts judged blind by an Opus reviewer; two pass adds a guided second round and a final blind rank against the carried-over round-one winner. Attempts can run on Anthropic models or GLM z.ai models (via the glm CLI)
3glm-pr-reviewerSet up an automated AI pull-request reviewer on Gitea or Forgejo Actions that runs Claude Code (headless) against a GLM model via the Z.AI Anthropic-compatible endpoint, posts the review as a PR comment, and sets a pass/fail status check that can gate merges. Use for AI code review in CI on self-hosted git servers, wiring Claude Code/GLM into Actions, or debugging an existing Claude-Code-on-Actions reviewer
4grill-meRelentless one-question-at-a-time Socratic interview that stress-tests a plan or design before any code is written. Maps the decision tree, resolves the highest-impact unknowns first with a recommended answer each, surfaces dependencies, pushes back on risky choices, and ends in a structured shared understanding. Planning-only. Use when someone says 'grill me' or wants to poke holes in / pressure-test / pre-mortem a plan
5mockoon-cliUse Mockoon CLI to create API proxies with full request/response logging, mock API servers for testing, and record real API interactions for debugging
6uv-python-managerStandardize Python development using UV - a fast, unified package and project manager