Security engineer builds a "World Map" knowledge base that keeps his agents hallucination-free
prasanna_says · x · 2026-11-22
Security engineer makash shares his agent architecture: a "Security World Map" knowledge base built from relationships between apps, APIs, and cloud/k8s infrastructure, with an extracted security knowledge base that humans haven't updated since April 26.
- Every agent he builds since treats the world map as its finite set of context, which he says eliminates hallucinations; the agent's open-ended "spec" derives from this map.
- He calls it "Context7 for security" and notes peers were skeptical until the vocabulary (Knowledge Factories) caught up.
- The post quotes nichochar's "Knowledge Factories" concept — curator agents attached to spaces ("gardeners") vs. agents attached to persons ("butlers" like OpenClaw, Grok Bot, Muse).
More from coding & agent
- FitMetrics runs Grok Build + 27 Grok Bots, merging 1,800 PRs in one month — XFreeze · 2026-11-22
- OpenAI ships 'Sign in with ChatGPT': local apps run GPT-6.x on your plan quota — alexcovo_eth · 2026-11-22
- The top programming languages of 2027: YAML, JSON, Markdown, and Bash — kmeanskaran · 2026-11-22
- Adversarial review makes AI agents flip-flop on every pass, user finds — EconomicsEvery4493 · 2026-11-22
- stonks MCP: open-source server that lets AI agents trade your financial instruments — QuixiAI · 2026-11-22
- Kyros Labs builds Kepler, one workspace to replace 5-tool robot policy debugging — Repulsive-Fee-2735 · 2026-11-22