mcp2agy is an MCP-native, multi-agent security audit engine engineered for the Google Antigravity (AGY) ecosystem.
Skill Claude CodeCodex
Code Fixer subagent definition for the mcp2agy pipeline. This agent produces minimal, correct, testable fixes for verified security findings. Do NOT use this agent directly — it is invoked by the audit pipeline orchestrator.
Skill Claude CodeCodex
Reporter subagent definition for the mcp2agy pipeline. This agent translates verified findings into publication-ready security reports and GHSA advisory drafts. Do NOT use this agent directly — it is invoked by the audit pipeline.
Skill Claude CodeCodex
Repository 0day Audit Scanner Subagent for the mcp2agy pipeline. Performs automated AST/data-flow triage and candidate discovery across 27+ vulnerability classes (Phases 0-2). Receives target files from the Orchestrator, grounds every hit in physical line counts and file hashes, outputs normalized candidates.yaml, and…
Skill Claude CodeCodex
Verifier subagent definition for the mcp2agy pipeline. This agent converts scanner candidate findings into rigorously confirmed vulnerabilities with deterministic PoCs. Do NOT use this agent directly — it is invoked by the audit pipeline orchestrator.
Skill Claude CodeCodex
Master 0day Audit Pipeline Orchestrator for Antigravity. Activated by /audit command. Orchestrates the complete multi-agent pipeline: Scanner → Verifier → Reporter → Fixer. Coordinates subagents using invokesubagent and sendmessage, manages structured workspace handoffs, enforces strict anti-hallucination integrity…
Skill Claude CodeCodex
Autonomous Multi-Agent CTF Solving Squad for Antigravity. Activated by /ctf command. Analyzes CTF challenges, automatically classifies problem domain (Pwn, Rev, Web, Crypto, Forensics), and spawns specialized solving subagents concurrently. Produces verified solve scripts, captured flags (flag{...}), and detailed…
Skill Claude CodeCodex
Comprehensive User Guide & Interactive Documentation for mcp2agy. Activated by /docs command. Provides complete architecture blueprints, slash command usage, MCP tool reference, multi-agent pipeline walkthroughs, workspace standards, cross-platform setup, and troubleshooting.
Skill Claude CodeCodex
Multi-Agent Exploit Development & Weaponization Engine for Antigravity. Activated by /exploit command. Orchestrates specialized exploit engineering subagents for memory corruption (ROP/Heap/UAF via pwntools & angr), web/API exploitation (SQLi/SSRF/SSTI/Deserialization), and multi-step exploit chain assembly. Generates…
Skill Claude CodeCodex
Security Code Remediation & Fix Engine for Antigravity. Activated by /fix command or invoked as the Fixer subagent by the audit orchestrator. Generates minimal, correct, testable fixes for verified 0day findings across 16 CWE classes. Enforces multi-site completeness, atomicity (no new TOCTOU), and produces regression…
Skill Claude CodeCodex
Security Tool Engineering & Fuzzing Forge for Antigravity. Activated by /forge command. Orchestrates subagents to engineer custom vulnerability scanners, coverage-guided fuzzing harnesses (AFL++, LibFuzzer, Boofuzz), detection rules (Sigma, YARA, Suricata), and high-performance offensive/defensive tooling in Python…
Skill Claude CodeCodex
Use when auditing for vulnerabilities similar to verified cases previously ingested by mcp2agy.
Skill Claude CodeCodex
Multi-Agent Reconnaissance & Attack Surface Intelligence Engine for Antigravity. Activated by /recon command. Orchestrates concurrent subagents for passive OSINT, active service discovery, web application mapping, API surface enumeration, and cloud asset discovery. Produces structured intelligence targets feeding…
Skill Claude CodeCodex
Security Report Composer & Advisory Specialist for Antigravity. Activated by /report command or invoked as the Reporter subagent by the audit orchestrator. Translates verified findings and synthesized exploit chains into publication-ready security advisories adhering to the standard impact-first format. Enforces…
Skill Claude CodeCodex
Multi-Agent Reverse Engineering & Binary Analysis Engine for Antigravity. Activated by /reverse command. Orchestrates specialized subagents for static decompilation (Ghidra/Radare2), dynamic binary instrumentation (Frida/GDB), firmware unpacking, and proprietary protocol reverse engineering. Produces annotated…
Skill Claude CodeCodex
0day Vulnerability Scanner for Antigravity. Activated by /scan command or invoked as the Scanner subagent by the audit orchestrator. Performs static triage, AST-level data flow analysis, and dangerous sink inspection across 27+ vulnerability classes. Produces normalized candidate records (candidates.yaml) with…
Skill Claude CodeCodex
Pipeline Status & Multi-Agent Workspace Monitor for Antigravity. Activated by /status command. Inspects active subagents (managesubagents list), detects host OS & platform parameters, tracks in-progress and completed audit runs, counts confirmed findings, verifies artifact inventories, and renders an executive status…
Skill Claude CodeCodex
Multi-Agent Coordinated Offensive Swarm for Antigravity. Activated by /swarm command. Launches a fully coordinated swarm of specialized subagents (Recon, Scanner, Verifier, Exploit, Reverse, Reporter) operating in parallel to perform an end-to-end security assessment on a complex target.
Skill Claude CodeCodex
Vulnerability Verifier & PoC Rigor Engine for Antigravity. Activated by /verify command or invoked as the Verifier subagent by the audit orchestrator. Converts candidate findings into confirmed 0days through 7-class confounder elimination, source-to-sink reachability proofs, deterministic PoC construction (≤ 60s…