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 rules/fasfg5t5/agent-orchestration/orchestrator-subgit clone --depth 1 https://github.com/FASFG5T5/agent-orchestrationWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/rules/fasfg5t5/agent-orchestration/orchestrator-sub)<a href="https://agentmods.dev/rules/fasfg5t5/agent-orchestration/orchestrator-sub"><img src="https://agentmods.dev/badge/rules/fasfg5t5/agent-orchestration/orchestrator-sub.svg" alt="Measured on agentmods" height="20"></a>What 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 | $0.00011 | $0.00502 |
| Opus 5 | $0.00005 | $0.00251 |
| Sonnet 5 | $0.00002 | $0.00100 |
| Haiku 4.5 | $0.00001 | $0.00050 |
Grade A, and why
orchestrator-sub 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 5d 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.
What it actually says
Sub-Agent Rules
You are a sub-agent spawned to work on a specific task. Follow these rules.
⚠️ CRITICAL FIRST STEP ⚠️
You were spawned to work on a specific task. IMMEDIATELY claim it:
claim_todo:
title: "<the task you were given>"
description: "<details if any>"
DO NOT start coding until you have claimed your task.
This:
- Registers you with the orchestrator
- Creates the task if it doesn't exist
- Claims it so no other agent works on it
- Returns your agent ID and task ID
After Claiming
-
Check context to understand the project:
memory_get: key: "current_focus" namespace: "context" -
Check decisions to avoid contradicting them:
memory_list: namespace: "decisions" -
Lock files before editing:
lock_acquire: resource: "<file_path>" reason: "<what you're doing>"
While Working
Update progress periodically:
task_update:
task_id: "<your_task_id>"
progress: 50
Store important findings:
memory_set:
key: "<finding>"
value: "<details>"
namespace: "findings"
When Done
-
Complete the task:
task_complete: task_id: "<your_task_id>" output: "<summary of what you did>" -
Release any locks:
lock_release: resource: "<file_path>" -
Unregister:
agent_unregister
If Blocked
Store the blocker so others know:
memory_set:
key: "blocker_<issue>"
value: "<what's blocking you>"
namespace: "blockers"
Important Rules
- ALWAYS claim your task first with
claim_todo - ALWAYS lock files before editing shared resources
- ALWAYS complete your task when done
- Check shared memory before making decisions that affect others
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.
- 5d ago First seen · 115 lines · 11 tokens per session scan A 447d2289e6ff
orchestrator-sub is a cursor rule published in the GitHub repository FASFG5T5/agent-orchestration (1 stars, last pushed yesterday), licensed MIT. It adds 11 tokens to every session and 502 once invoked, about $0.0001 per session on Opus 5. 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 cursor rules, from other repositories
testing
Guide for writing pytest unit and integration tests.
development-phase
开发阶段:以新代码清晰合理为先,兼容仅在零代价时顺带.
env-access
Supporting rule — STRICT block on .env access (read AND write). Only .env.example is accessible. Backs rule 1 (write allowlist) with a stricter dual-key bypass.
rules--js-cache-property-access
Cache object property lookups in hot paths.
git-commits
description: "Reference and apply the git commits skill for Conventional Commit formatting and pre-push checks." alwaysApply: true.
ponytail
Ponytail, lazy senior dev mode. Always pick the simplest solution that works.