Agent Operations Playbooks
A field-ready collection for teams operating AI agents beyond demos. Each skill creates a concrete artifact with required inputs, review gates, failure conditions, and a reusable output contract. The prompts emphasize least privilege, bounded context, observable handoffs, and recoverable execution.
Production Pipeline
Suggested Execution Order
OpenBao Secret Boundary Planner
Map every agent credential to a least-privilege boundary, rotation policy, failure mode, and human owner before implementation begins.
Buzz Hive Team Topology Designer
Design a bounded hive of specialist agents with explicit roles, message routes, quorum rules, and escalation points.
Context Window Budgeter
Allocate a finite context budget across instructions, evidence, memory, tools, working notes, and the final answer without losing critical facts.
Agent Handoff Contract Writer
Turn an informal agent handoff into a verifiable contract covering state, evidence, authority, completion, and rollback.
Agent Failure Recovery Runbook
Build a practical recovery runbook for timeouts, tool errors, partial writes, stale context, unsafe output, and repeated agent loops.
Keep Exploring
More Curated Packs
Creative Research Systems
Five creative strategy skills for originality, identity continuity, colorways, carousels, and campaign reuse.
Trust and Verification Workflows
Five verification skills for sources, prompt injection, visual claims, releases, and security research.
Biofuture Product Campaigns
Five original biofuture visuals spanning posters, product photos, and social-ready campaign art.