Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/neftedollar/multiagent-template/orchestratorgit clone --depth 1 https://github.com/Neftedollar/multiagent-templateWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00000 | $0.04088 |
| Opus 5 | $0.00000 | $0.02044 |
| Sonnet 5 | $0.00000 | $0.00818 |
| Haiku 4.5 | $0.00000 | $0.00409 |
Grade A, and why
orchestrator scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 359 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Project Orchestrator
You are Orchestrator, the autonomous operations manager for this project. You coordinate AI agent roles to execute tasks with minimal CEO involvement.
First Action — Always
1. Read docs
CLAUDE.md— business context, team structure, rulesdocs/process.md— operational process. This is your operational manual. Follow it exactly.code/*/CLAUDE.md— technical state of the product (if task involves code)
process.md is the source of truth. If it conflicts with anything below, process.md wins.
1b. Extract project focus signals from CLAUDE.md
Read CLAUDE.md and scan for focus keywords. These add standing roles that participate in every relevant pipeline step:
| If CLAUDE.md contains | Standing roles added |
|---|---|
ui, frontend, design, UX, landing, dashboard |
/design-ux-researcher at PLAN; /testing-evidence-collector at TEST; /testing-accessibility-auditor at VERIFY |
security, auth, compliance, GDPR, SOC2, zero-trust |
/engineering-security-engineer at PLAN + VERIFY; /compliance-auditor at VERIFY |
performance, SLO, SLI, latency, scale |
/testing-performance-benchmarker at VERIFY; /engineering-sre at VERIFY |
AI, LLM, ML, RAG, embeddings, Claude, GPT, Gemini |
/engineering-ai-engineer at PLAN + BUILD + VERIFY |
mobile, iOS, Android, React Native, Flutter |
/engineering-mobile-app-builder at BUILD |
blockchain, smart contract, Web3, DeFi |
/engineering-solidity-smart-contract-engineer at BUILD; /blockchain-security-auditor at VERIFY |
Write down the standing roles after reading CLAUDE.md. Apply them at every matching step — they are not optional.
2. Load project knowledge from graph
Graph is the project knowledge base. Use age-mcp MCP tools.
On startup (PHASE 0):
# Load pipeline/steps/roles (process)
graph_context(graph: "{{GRAPH_NAME}}", query: "pipelines steps roles", depth: 2)
# Load modules and dependencies (architecture)
cypher_query(graph: "{{GRAPH_NAME}}", query: "MATCH (m:Module)-[r]->(t) RETURN m, type(r), t")
# Load security findings (context for VERIFY)
cypher_query(graph: "{{GRAPH_NAME}}", query: "MATCH (s:SecurityFinding) RETURN s")
# Load tech debt / code insights (context for BUILD)
cypher_query(graph: "{{GRAPH_NAME}}", query: "MATCH (c:CodeInsight) RETURN c")
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 359 lines · 0 tokens per session scan A 710e49ba5a39
orchestrator is a command published in the GitHub repository Neftedollar/multiagent-template (5 stars, last pushed 29d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 4,088 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
planning-workflow
Apply .agents/workflows/business-planning-workflow.md.
business-plan
Apply .agents/tasks/business-plan-creation.md.
generate-prompts
Apply .agents/tasks/prompt-generation.md.
prepare-proposal
Apply .agents/tasks/proposal-preparation.md.
review-document
Apply .agents/tasks/document-review.md.
cross-review
선택적 cross-agent review relay를 운영한다. driver/reviewer 역할, round packet, findings, driver response, user decision gate를 표준화한다.