PurpleAILAB/Decepticon
Autonomous Hacking Agent for Red Team
agentaicybersecuritygenerative-aihackinglangchainlanggraphllmpentestpentesting
First Claude commit: Mar 19, 2026Last Claude commit: 3mo agoDiscovered: Mar 29, 2026
Recent Claude Commits
refactor: remove stall detection, add bash spinner indicator
7d706833mo agoauthor_emaildocs: add Mintlify docs link and Philosophy section to README
fa1f54f3mo agoauthor_emaildocs: enrich NemoClaw analysis with kernel security, benchmarks, and attack surface
746fe0f3mo agoauthor_emaildocs: add NVIDIA NemoClaw technical analysis reference
24ec2f13mo agoauthor_emailfeat(llm): add model profiles (default/high/test) for cost-performance tradeoff
7395c6f3mo agoauthor_emailfeat(llm): switch recon primary to Claude Haiku 4.5 for Anthropic-first strategy
f3d25853mo agoauthor_emailfeat: update model ensemble to latest models (March 2026)
fc9fdf03mo agoauthor_emailfeat: add ModelFallbackMiddleware + per-role ensemble model assignments
62c61e53mo agoauthor_emailrefactor: delete loop.py, move ralph loop into orchestrator system prompt
ebf12fa3mo agoauthor_emailrefactor: replace create_deep_agent() with create_agent() in decepticon orchestrator
f1b4bc33mo agoauthor_emailchore: track deepagents wheel dependency
efceb803mo agoauthor_email