no_human
Open-source agent that takes a ticket to a reviewed pull request, with proof
Emerged finds new products worth a look, daily — what’s new today
captain-hook
Turns the corrections you keep giving Claude Code into typed, tested hooks
It mines your Claude Code transcripts for the things you keep telling the agent not to do and opens a PR enforcing each one as a Python hook, so the rule lives in code, not in a prompt.
captain-hook is a hooks framework for Claude Code from Yasyf Mohamedali, a three-time founder, introduced by an essay arguing that constraints beat longer prompts. Hooks are typed Python functions with tests: the stock ones block force-pushes and rm -rf, flag print statements, and gate a task as done until tests pass. Each returns a hint, substitution or warning that lands back in the model's context. The distinctive part is a session reviewer that mines transcripts for repeated corrections and proposes new hooks as PRs. Brew, uvx or plugin install. PolyForm Noncommercial, 13 stars.
More like this
JARVIS
A voice assistant that brainstorms with you, then runs and watches Claude Code
Unlazy
An agent skill that refuses to call work done until executable gates pass
DashClaw
Approval gates and an audit ledger for coding agents left running unattended
TokenEconomizer
Cuts what Claude Code and Codex spend in a repo you work in daily
Stroq
A local firewall that stops AI agents acting on hostile content they just read