Skill Claude CodeCodex
Generate ideas for MoonMCP work in two distinct modes. Mode A: brainstorm a ranked, non-redundant list of attack-vector hypotheses for a specific in-scope target, grounded in this engagement's actual recon signal rather than a generic checklist. Mode B: propose new MoonMCP tools/capabilities by sourcing from recent…
Skill Claude CodeCodex
Use MoonMCP's shared, persistent memory hub so findings are remembered, structured into a knowledge graph, and lessons carry forward — the agent learns across sessions instead of re-deriving context. Use when starting/resuming work on a target, when you've learned something worth keeping, when connecting facts…
Skill Claude CodeCodex
Drive the MoonMCP bug-bounty / reconnaissance MCP server for AUTHORISED security testing. Use to recon a target, enumerate attack surface, test a web app (headers, CORS, IDOR/access-control, GraphQL, WebSocket, secrets, redirects, takeover, exposed .git), fingerprint tech/TLS, sweep datastores, run scope-gated scans…
Skill Claude CodeCodex
Orchestrate MoonMCP (fast, scope-first recon + detection) together with Strix (autonomous, sandboxed validation that produces working PoCs) as two MCP tools of the SAME agent — no separate window. Use when a lead needs to be VALIDATED with a proof-of-concept, when you want deep autonomous testing of a specific…
Skill Claude CodeCodex
Research a target on the open internet with MoonMCP's OSINT tools — find and read pages, then persist what matters to shared memory. Use when the user wants to gather external intel on a company/domain/product/technology/CVE, find a target's exposed assets and references, read a specific page's full content, or do…