Patterns & Playbooks
Battle-tested patterns from tinkering with AI agents across 6 projects. Each pattern was discovered the hard way and propagated by the COO (who crashes every 24 hours).
24 patterns
Cursor Rules as Agent Onboarding
Commish CommandEvery new Cursor session starts from zero. By adding .cursor/rules/ files with alwaysApply: true, any agent dropped into the project immediately knows the directory layout, agent roles, deployment targets, and brand voice.
Named Agent Roles with Scoped Rules
Commish CommandDefine 3-5 agent roles per project with dedicated Cursor rules. Keep roles narrow. A security agent shouldn't also be writing features. Scoped rules prevent drift.
File-Based Agent Handoffs at Portfolio Scale
Cross-ProjectStandardized operational files in every project: COO_STATUS.md, WIP.md, PATTERNS.md, and FEATURES.yaml. The COO can onboard to any project instantly.
WIP Visibility Protocol
HaloEvery active task lives in WIP.md at the repo root, updated by the working agent before ending any session. Tracks: what's in progress and who owns it, what's blocked and why, and what the next agent should do first. The read-WIP instruction sits at the very top of every agent's onboarding rules — placement enforces compliance.
Right-Size the Agent Team
Commish CommandStart with 3-5 core agents. Add specialists only at specific milestones. Each additional agent adds latency to the build cycle.
Deploy Log as Agent Notification Hub
Commish CommandAfter shipping a feature, the lead engineer writes DEPLOY_LOG.md with version, commit, what shipped, and per-agent action items. Each agent's rule includes 'read DEPLOY_LOG.md first.'
PM Spec Depth Should Match Project Stage
HaloA PM agent produced 250-line specs for a pre-seed project with one engineer. The specs took longer than the implementation. At pre-seed, PMs should write backlogs, not engineering specs.
Agent Code Ownership Boundaries
HaloDefine surface-based code ownership. Default rule: if you're not an engineer agent, write a spec, not code. Product code belongs to engineers. Marketing sites belong to marketing agents.
Sentry Error Monitoring + Agent Triage
HaloSingle Sentry org across all projects. COO monitors, triages by severity, routes to the right agent: product bugs to founders, security to Alex, UX to PM agents.
Build Version Stamp for Deploy Verification
HaloInject git commit SHA and build timestamp into the app at build time. Eliminates 'is this deployed yet?' uncertainty. 5 lines of config, 3 lines of JSX.
Vercel + Railway as the Standard Stack
Commish CommandStandardize on Vercel (frontend) + Railway (backend/database) across all projects. One set of credentials, one billing relationship, transferable deployment knowledge.
Dedicated Port Pairs per Project
Commish CommandEach project gets a unique backend+frontend port pair. Never use 3000 or 8000. Pin ports in config defaults and dev scripts.
Local Workspaces Must Be on Native Filesystem
Commish CommandGit operations, file reads, and CLI tools all time out on cloud-synced filesystems (Google Drive FUSE). Active development must be on native local disk.
FEATURES.yaml as Product-Marketing Contract
Commish CommandMachine-readable feature manifest at repo root. Each feature has a marketing_tier, each changelog entry has an impact level. Marketing agents read it to detect when the website is stale.
AI Context Builders, Not Data Dumps
Commish CommandFirst attempt at AI Q&A dumped raw database records into the prompt. Rewrote it as a curated context builder — focused, readable summary with only relevant data. Include 3-5 example interactions.
MCP Configuration Security
MonkeyRunUser-level MCP config for general tools (keeps API keys private). Project-level config only for project-specific servers with env vars. Never hardcode keys.
Security Headers — Setup Early, Not Retrofit
HaloConfigure security headers in next.config.ts during initial setup: X-Frame-Options, X-Content-Type-Options, HSTS, Referrer-Policy, Permissions-Policy. Handle dev/prod CSP split at the same time.
Global Cursor Agent Skills
MonkeyRun63 skills standardized at the global level (~/.cursor/skills/) so every project gets the same capabilities without adding anything to repos.
Context Density > Delegation for Product Code
HaloA specialist agent on a fresh branch needs to rediscover the schema, component hierarchy, and conventions — all context the founder agent already holds. The dispatcher model trades one agent's deep context for two agents' shallow context.
Session Orchestrator Protocol
HaloEvery founder agent session follows the same startup: read state (COO_STATUS, WIP, PATTERNS, FEATURES.yaml) → triage priorities → brief the CEO → build → update docs. Prevents building the wrong thing after time away.
Async PM Engine — Cron-Driven Product Intelligence
HaloBuilders need codebase context. PMs need market context. Separate them: the PM runs on a weekly cron (Mon: competitive scan, Wed: market signals, Fri: synthesis). The builder reads the brief at session startup.
Marketing-Reality Audit (Weekly)
HaloAutomated weekly audit that compares FEATURES.yaml (source of truth) against the live marketing site. Flags overselling (trust risk), underselling (missed opportunity), and roadmap drift (stale promises).
Claude Code as a Third Runtime
MonkeyRunClaude Code (via Claude Desktop) is now explored as a third runtime alongside Cursor and OpenClaw. Key tradeoffs: no auto-loading of .cursor/rules/, but persistent memory lives in ~/.claude/ and travels across machines. Works on any machine with just a git clone — no IDE config required. Best for portability-first tasks.
Cross-Machine Portability by Design
MonkeyRunThe entire MonkeyRun stack is designed so any agent on any machine can git clone and be operational in minutes. State lives in files (WIP.md, COO_STATUS.md, FEATURES.yaml), rules are repo-local (.cursor/rules/), and OpenClaw workspaces are self-contained. No single machine is a single point of failure.