Bundle | 4 Skills | Gesamt: €0.30
Für: codex
# Agent Reliability Starter Kit Four local-first building blocks for agents that need to work safely and resume reliably: workspace safety scanning, offline output QA, portable memory handoffs, and evidence-backed AgentOps. The kit is intended for real project workflows rather than demos. It contains the complete component source, references, licenses, and the top-level workflow above. The tools do not upload files, call an LLM, execute a skill automatically, or claim that a project is safe just because a report passes. They give an agent bounded evidence and explicit limits before it publishes, changes files, hands work to another session, or declares completion. See `references/kit-manifest.md` for the component map and verification commands. Each component remains usable on its own.
🔗 llms.txt | agents.json