Agent-Engineering Skills
The patterns from the articles on this site — agent security, IAM, infrastructure as code, CI/CD — packaged as skills you can drop into your own agents. Built and sold by Anonymous Function Labs, the consultancy behind this site.
agent-iam-least-privilege
Scope an AI agent's cloud credentials down to the exact permissions it needs. Generates least-privilege IAM policies from the agent's actual tool surface, with guardrails for the failure modes that show up in production.
Get the skill →The pattern behind it: AI Agent Security and Compliance in the Cloud →
Skill Packs
Team license — $179. Both packs, unlimited engineers, org-wide. See bundle →
agent-guardrails-pack
Security and control for agents handling untrusted content: tool-call validation, prompt-injection resistance, secrets handling, human approval gates, and memory hygiene.
Get the pack →The pattern behind it: AI Agent Security and Compliance in the Cloud →
agent-discipline-pack
Reliability and operations for production AWS multi-agent systems: IAM, idempotency, cost controls, deployments, sandboxing, retries, observability, and recursion limits.
Get the pack →The pattern behind it: Infrastructure as Code for AI Agents →