qmilab/lodestar
The trust layer for AI agents.
The trust layer for AI agents.
npx add-skill qmilab/lodestarThe trust layer for AI agents.
Agent-side preflight skill for vaultpilot-mcp — MCP-independent integrity checks loaded by Claude Code from ~/.claude/skills/
Catches the bugs linters miss — logic-first code review via execution tracing. Finds behavioral bugs, type-contract breaches & semantic mismatches across 9 risk categories. Six skills: review, explain, diff, locate, health-check, and autonomous audit-and-fix pipeline. Claude Code · Codex CLI · Gemini CLI.
Agent Skills
OAuth token broker (apl CLI) + Claude/Codex agent skill for Google Workspace, Microsoft 365, and GitHub
Open source Claude Code plugin — 52 invisible engineering skills turn vibe coding into production-ready software. Auto specs, TDD, security hooks, quality gates, self-improving retrospective. 97% vs 41% on blind benchmark.
Agent Skill for Claude: Design, build, and debug multi-agent swarms. 8 archetypes, 5 patterns, deterministic-first cascade, chalkboard protocol.