TurboFlow Prompt Guide
Description
> Quick reference. Your CLAUDE.md has the full details — this is what you actually type. ---
Installation
claude install-skill https://github.com/marcuspat/turbo-flow README
TurboFlow Prompt Guide
Quick reference. Your CLAUDE.md has the full details — this is what you actually type.
1. START A SESSION
Paste this as your first message every time:
Boot up. Project ID is [your-project-name]. Run the full session boot protocol, show me the status board when done.
That's it. The CLAUDE.md runs the full 11-step primer automatically:
Step 0: Pre-flight — env file, migrations, stashes, disk space
Step 1: Daemon — starts background workers
Step 2: Hooks — verifies registration, fires session-start (restores context from SQLite)
Step 3: Beads — loads all open issues, blockers, decisions
Step 4: Intelligence — checks if warm, auto-runs pretrain if cold
Step 5: Ruflo Memory — verifies HNSW index health, then recalls project state
Step 6: AgentDB — health check probe via MCP, auto-repairs if unavailable
Step 7: GitNexus — checks freshness, auto-reindexes if stale (not optional)
Step 8: Swarm — initializes star topology
Step 9: Security — verifies pre-edit hooks and audit worker are active
Step 10: Router — routes today's work via Q-learning
Every subsystem gets primed, not just checked. If anything is broken, it auto-fixes and reports what it did.
If resuming a previous session:
Boot up. Project ID is [your-project-name]. Resume from last session — check memory and beads for where we left off.
After boot: Full branch analysis
Once the status board is green, paste this to get a complete picture of where the branch stands and what needs to happen:
Run a full branch analysis. I need to know exactly where we are and what's left to do.
1. BRANCH STATE — what branch, how far ahead/behind main, full diff summary vs main, uncommitted changes, stashes, last 10 commits with timestamps
2. CODEBASE CHANGES — run gitnexus_detect_changes vs main, what symbols changed, what execution flows are affected, any HIGH/CRITICAL risk
3. BEADS STATUS — open issues, blockers, recently completed, anything flagged for human decision, run bd lint
4. MEMORY RECALL — search memory for recent sessions, known blockers, AgentDB patterns related to current work, logged decisions
5. HEALTH CHECK — run tests, build, and lint right now. Check for pending migrations and missing env vars.
6. QUALITY ENGINEERING — run aqe-gate on the branch. Show test coverage gaps, security scan results, and any quality issues found. If tests are missing for changed files, flag them.
7. GAP ANALYSIS — based on everything above, what is DONE, IN PROGRESS, and NOT STARTED? What's blocking? What can be parallelized?
Then give me:
A. STATUS SUMMARY — 5-10 lines, plain English
B. PRIORITIZED TODO — generate a TodoWrite ordered by priority and dependency, include any aqe-gate findings as tasks
C. SESSION PLAN — if I have 2-4 hours, what do we tackle first?
This runs ~20-25 tool calls across git, gitnexus, beads, memory, agentdb, agenti
Related Agents
Openai Codex CLI
(55.8k ⭐) - Lightweight coding agent that runs in your terminal.
Development community src/agents/ — 11 Agent Definitions
**Generated:** 2026-04-11
Development community Cavecrew Builder
>
Development community Cavecrew Investigator
>
Development community Cavecrew Reviewer
>
Development community Contributing to nanobot
Thank you for being here. nanobot is built with a simple belief: good tools should feel calm, clear, and humane. We care deeply about useful features, but we also believe in achieving more with less:
Development community