I build the infrastructure around coding agents: skills, trust, memory, orchestration, routing, governance, and recovery.
Each project works alone. Together, they form a closed execution loop.
Choose the problem you want to solve, then follow that project's setup guide.
Each repository documents its own installation, requirements, support, and license. The stack diagram below describes how the projects relate; it is not a shared installer.
// OPEN_MODULE_BAY — more systems, tools, and experiments
| Project | Role |
|---|---|
awesome-goal-prompts |
114 source-backed /goal contracts for coding agents |
argus |
Install-time supply-chain scanner for npm, PyPI, and crates.io |
specrail |
Spec-first rails for agent-assisted repository workflows (archived) |
keepline |
Session command center for monitoring and recovering agent work |
| Project | Role |
|---|---|
sage |
Blazing-fast coding agent in pure Rust |
rnk |
Declarative TUI framework with React-like hooks and 45+ components |
rui |
GPU-accelerated UI framework inspired by GPUI |
ccstats |
Claude Code and Codex token/cost analytics CLI |
jsonrepair-rs |
Repair 30+ classes of malformed JSON from LLM output |
rekey |
Agent API-key routing and credential-isolation proxy |
rclean |
Find and clean rebuildable developer artifacts |
| Project | Role |
|---|---|
tokenpulse |
Read-only Codex and Claude Desktop throughput analysis with consistent timing and a controlled benchmark protocol |
loom |
Skill registry and projection control plane (no new features) |
claude-skill-manager |
Discover, install, and manage Claude Code skills |
claude-skill-registry-core |
Deduplicated registry artifacts and index |
claude-skill-registry-data |
Raw skills archive data |
auto-contributor |
Automated GitHub contribution workflow powered by Claude Code |
cc-model-watch |
Warn when Claude Code silently swaps the serving model |
ccp |
Isolated Claude Code profiles — run multiple API providers side by side |
claude-code-anime-sounds |
Anime-themed Claude Code hook sounds |
| Project | Role |
|---|---|
refine |
Extract reusable knowledge from coding-agent conversations |
chat-archive-rs |
Archive and search Claude Code and Codex sessions |
stash |
Evidence-backed personal task inbox |
open-source-repo-ledger |
Public repository-readiness ledger |
shipwise |
Agent-facing open-source launch playbook |
test-loop |
Language-agnostic drift detection and CI guardrails |
seo-agent-suite |
Discoverability audits for Shipwise |





