Agent Ops Playbook Pro

What This Is

The Agent Ops Playbook — but production-hardened. Built from real failures, not theory. Covers everything the basic playbook misses: write-through enforcement, multi-project isolation, sub-agent handoff discipline, and a validator script that catches memory drift before it costs you.

What Makes It Different

Every other ops guide gives you templates. This one gives you enforcement:

• Memory sync validator script (runs every heartbeat, blocks commits when stale)
• Pre-commit hook that prevents pushing with out-of-sync project files
• 6-rule write-through system: what to write, when, and which file
• Real post-mortems from 4 production failures with root cause and rule created

Multi-Project Isolation

Running multiple projects simultaneously across different Telegram groups or channels? This covers the full pattern: index.json routing, per-project TASKS.md + MEMORY.md, context isolation, staleness checks, and cross-project work in the main session.

What Is Included

• Full SKILL.md playbook (session discipline, memory architecture, heartbeat/cron patterns, safety, comms)
• Starter templates: AGENTS.md, SOUL.md, USER.md, HEARTBEAT.md, TASKS.md, MEMORY.md, index.json
• Memory architecture deep-dive: three layers, multi-project isolation, enforcement layers
• 4 failure post-mortems: credentials lost, exec lockout, provider shutdown, memory drift
• memory-sync.sh validator script — drop in and run