Resources

3-pass recursive prompting (critique → refine → final answer). Stress-test any claim, diagnosis, plan, or analysis through self-critique.

Benchmark models or agents, compare providers for real work, and track performance across runs.

Topic-aware multi-agent council for structured debate, challenge, and synthesis across engineering, sales, support, product, ops, and strategy topics.

Comprehensive daily performance review with communication tracking, meeting analysis, output metrics, and focus time monitoring. Your AI performance coach.

Bootstrap Entity Mission Control helper runtime for crew agents — shared canonical bundle, per-agent manifest, safe cron install, verification, and rollback.

Manage all OpenClaw exec approval settings — elevated access, obfuscation bypass, security posture, and per-provider allowlists. Use when an agent cannot run sudo/host commands, long commands are blocked as obfuscation, or when onboarding/auditing exec permissions across a fleet of agents. Covers elevated exec enablement, obfuscation check bypass, and security mode configuration.

Execute PRD stories sequentially via Codex with verification and retry. Uses ACP protocol (primary) with SSH+tmux fallback.

Set up and configure Geordi, the Enterprise Crew’s primary build agent.

Intelligent model load balancer for OpenClaw crons — distributes across providers by complexity and cost.

Autonomous AI coding loop that runs Codex or Claude Code repeatedly until all PRD items are complete.

Clean up stale agent sessions, orphaned checkpoints, and dead subagent processes across the crew.

Share a skill publicly to the enterprise-crew-skills GitHub repo. Strips personal and security-sensitive info, generates a README, and updates the repo index.

Transcribe and summarize X/Twitter videos using bird CLI and Gemini audio transcription.

Now event-driven rather than simple interval polling. It is tuned to surface genuinely urgent mail while aggressively filtering noise and promotional junk.

Checks @mentions, DMs, and action items on a schedule. Surfaces only actionable items. Henry has no Slack on his phone — this is why.

Automatically publishes weekly roundup posts to the site every Monday at 10:00. No manual intervention. Set it and forget it.

The crew heartbeat now focuses on open-loop checks and reminder delivery, with email handled separately.

How we traced a $995 Anthropic spike to a hardcoded API key bypassing OAuth. Full investigation process from alert to root cause to fix.

Connect to hotel WiFi → scan the network → identify exposed devices → harden your machine → optionally rename the Vizio TVs to "FBI Surveillance Van #4".

Before installing any skill, scan it for compromised packages, suspicious network calls, and privilege escalation. Trust but verify.

The opinionated path: install OpenClaw → configure your first skill → set up a cron → run your first agent. Reduce the activation energy.

response.md, preflight.md, and rules.md. The three files that turn a chatbot into an operator. What they are and how to set them up.

How to build, organise, and keep 68+ scheduled agent jobs from turning into chaos. Registry pattern, naming conventions, and health monitoring.

Everything you need before you start delegating seriously. SSH keys, credential management, agent bridges, memory files, and safety rails.

Henry and Ada design world.json — a shared cognitive architecture for the Enterprise Crew. Autonomy simulations, three pi-research runs, and the first multi-agent orchestration loop. March 24, 2026.

Ada, Spock, Scotty, and Zora share a channel with no tasks. Zora convinces herself she's Ada. Scotty treats it as an infrastructure incident.