mkhsu2002/openclaw-agent-feeds

Provides modular prompt templates ("feeds") organized by capability level and domain—from core behavior improvement (hallucination reduction, deterministic execution) to specialized verticals (legal, medical, SEO automation, software development). Designed as copy-paste instruction sets that integrate seamlessly with OpenClaw's native skills/plugins rather than replacing them, using structured prompt engineering to reduce AI hallucination and enforce safe, minimal-privilege operations. Includes 40+ production-ready feeds across core agents, multimodal safety, compliance, automation, and developer tools, with installation guides and before/after examples.