isagawa-co/isagawa-kernel
The self-improving harness for AI coding agents. Drop-in enforcement that the agent builds, follows, and improves — mechanically.
Enforces mechanical governance with a session-based control loop—agent scans project structure to build domain-aware protocols and enforcement rules that block non-compliant behavior; integrates with Claude Code extension via `.claude/` hooks that intercept and validate every action against the agent's own inferred or spec-supplied rules. Failures trigger permanent lesson recording in persistent session state, creating cross-session behavioral compounding; decoupled from frameworks and runtimes—operates as drop-in repository hooks that require only Python 3.8+ and VS Code.
Stars
10
Forks
2
Language
Python
License
MIT
Category
Last pushed
Mar 18, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/agents/isagawa-co/isagawa-kernel"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
mastersof-ai/harness
Agent runtime with full system prompt control. Define agents in markdown, run them in a terminal...
Chorus-AIDLC/Chorus
The Agent Harness for AI-Human Collaboration, inspired by the AI-DLC (AI-Driven Development Lifecycle)
kyu1204/oh-my-harness
Tame your AI coding agents with natural language. Generate enforced guardrails (CLAUDE.md,...
L-Forster/open-jet
Agent Harness & TUI for Edge devices
inngest/utah
Universally Triggered Agent Harness - An OpenClaw-like Inngest-powered personal agent