Meta-orchestrator that classifies tasks and routes to the right combination of installed skills across caveman (output compression), superpowers (dev workflow), anthropic-skills (business/domain expertise), and wired tactical libraries (wshobson, davila7, impeccable, anthropic-official, rea). 27 routing categories including whole-codebase analysis via repomix-pack, LLM/AI app dev, reverse-engineering/binary analysis via REA + ida-pro-mcp, live DAP debugging (debug-skill), an optional opt-in scientific-skills library (K-Dense, bio/chem/med), a live task dashboard, meta skill discovery, and curated tactical skill libraries (wshobson, davila7) for backend/security/CI-CD/database depth. Also scaffolds the canonical project-doc spine (README/AGENTS.md/CLAUDE.md/PROJECTLOG.md) in new projects via `/agent-master scaffold`. Invoke with /agent-master. Use as default entry point for ambiguous or multi-domain requests.
Security for LLM and agent systems: direct and indirect prompt injection, the OWASP Top 10 for LLM Applications, MITRE ATLAS technique mapping, excessive agency and tool-scope containment, egress allowlisting to stop data exfiltration, human confirmation for irreversible actions, and treating third-party skills and MCP servers as untrusted code. Use when an agent is given tools or credentials, when retrieved documents or repository files could carry injected instructions, or when reviewing an AI feature before it reaches production.
Pre-publish check for the engagement-review skill — run before any git commit or push to GitHub, so no client data, name, figure or identifier reaches the public repository.
How to use the spitfire-storybook skills in this repo — generate a personalized development book (PDF/EPUB) and narrate it into an audiobook (.m4b). Apply when the user asks to build a personalized book, run the book-builder, or narrate text/markdown/a book into audio.
Surya8991Cursor rule
