subagent-orchestrator

subagent-orchestrator is a cursor rule for Cursor from kok-o/koko-contextos-agents. It costs 27 tokens per session (853 once invoked), scanned A, original, MIT.

A method for coordinating multiple coding agents by splitting work into separate tasks, giving each agent limited context, and combining their results. It prevents agents from editing the same files at the same time.

In plain words
What is it for?
It helps divide work across modules, services, test suites, research tasks, security reviews, and documentation work.
Why use it?
It makes large or parallelizable work easier to organize and reduces conflicts when separate agents work on different parts of a project.

Cursor rule for Cursor

Written for Cursor: installed under .cursor/. Also seen: mentions subagents; built for gstack.

Good fit It helps divide work across modules, services, test suites, research tasks, security reviews, and documentation work.

Compare 6 cursor rules from other repositories ↓
Install with agentmods
npx agentmods add rules/kok-o/koko-contextos-agents/subagent-orchestrator
Install

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.

Clone the repo
git clone --depth 1 https://github.com/kok-o/koko-contextos-agents

Made for: Cursor.

Wrote 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.

agentmods badge for subagent-orchestrator

README.md
[![agentmods](https://agentmods.dev/badge/rules/kok-o/koko-contextos-agents/subagent-orchestrator/github.svg)](https://agentmods.dev/rules/kok-o/koko-contextos-agents/subagent-orchestrator)
Your own site
<a href="https://agentmods.dev/rules/kok-o/koko-contextos-agents/subagent-orchestrator"><img src="https://agentmods.dev/badge/rules/kok-o/koko-contextos-agents/subagent-orchestrator/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for subagent-orchestrator

Your own site · 80×15
<a href="https://agentmods.dev/rules/kok-o/koko-contextos-agents/subagent-orchestrator"><img src="https://agentmods.dev/badge/rules/kok-o/koko-contextos-agents/subagent-orchestrator.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 27 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 853 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00027 $0.00853
Opus 5 $0.00014 $0.00426
Sonnet 5 $0.00005 $0.00171
Haiku 4.5 $0.00003 $0.00085

Measured 4d ago against content hash f6e471a0ce50, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

subagent-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 4d 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.

.cursor/rules/subagent-orchestrator.mdc · 103 lines

How it starts

The opening of the file, as written. The whole thing — 103 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Skill: subagent-orchestrator

subagent-orchestrator

Overview

Multi-agent coordination protocol inspired by obra/superpowers. Enables a primary orchestrating agent to decompose complex workflows into isolated, parallel sub-tasks, delegate them with precise context boundaries, monitor execution, and synthesize outputs with zero merge conflicts.

When to Use

Activate whenever:

  • A task can be parallelized across distinct modules, services, or test suites.
  • Long-running exploratory research or multi-file refactoring exceeds single-context budget.
  • Running autonomous subagent workers for specialized roles (e.g. specialized QA tester, Security auditor, Docs generator).

Rules & Patterns

1. The Blast Radius Boundary Rule

Before delegating any subagent task:

  • Zero File Overlap: Each subagent MUST have a mutually exclusive list of target files. Two subagents must never be instructed to edit the same file concurrently.
  • Explicit Inputs & Outputs: Provide only the minimal schema, contract, or mock that the subagent needs. Do not dump the entire workspace into subagent prompts.

2. The 4-Step Delegation Lifecycle

[ Orchestrator ]
       │
       ├─▶ 1. DECOMPOSE: Break into orthogonal tasks with non-overlapping file sets
       │
       ├─▶ 2. DISPATCH: Launch subagent with precise goal, constraints, and finish criteria
       │
       ├─▶ 3. AWAIT & VERIFY: Validate subagent output against its individual quality gate
       │
       └─▶ 4. SYNTHESIZE: Merge subagent results into the main branch and run global regression suite

3. Context Hand-off Specification

Every subagent dispatch prompt must contain:

  1. Target Objective: Single, verifiable deliverable.
  2. Read-Only Context: Files to consult as reference without modifying.
  3. Write Scope: Exact file paths the subagent is permitted to create or modify.
  4. Completion Signal: Explicit instruction to report DONE with test evidence or BLOCKED with reason.

Read the full file on GitHub · 103 lines

Changes

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.

  1. 4d ago First seen · 103 lines · 27 tokens per session scan A f6e471a0ce50

Subscribe to this mod's changes

subagent-orchestrator is a cursor rule published in the GitHub repository kok-o/koko-contextos-agents (2 stars, last pushed 5d ago), licensed MIT. It adds 27 tokens to every session and 853 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 2026-09-05.

Related

Other cursor rules, from other repositories

deepen-architecture

Find deepening opportunities in a codebase, informed by the domain language in specs/tech-architecture/tech-stack.md and the decisions in specs/adr/. Use when the user wants to improve architecture, find refactoring opportunities, consolidate tightly-coupled modules, or make a codebase more testable and AI-navigable.

danielvm-git/bigpowers · 67 tokens

craft-skill

Create new bigpowers skills with proper structure, progressive disclosure, and bundled resources. Use when user wants to create, write, or build a new skill for the bigpowers lifecycle.

danielvm-git/bigpowers · 36 tokens

grill-me

Interactive assumption-surfacing Q&A that stress-tests a plan through relentless questioning until every decision is resolved. Use when user wants to challenge a plan, validate decisions from conversation/context, or mentions "grill me". For doc-grounded variant, use grill-with-docs.

danielvm-git/bigpowers · 55 tokens

run-planning

DISCOVER-PHASE ADVANCER — Drive the discover-phase checklist (specs/planning-status.yaml) through survey-context → scope-work → research-first → elaborate-spec → plan-release → slice-tasks. NOT a duplicate of plan-work or the planning spine; it orchestrates the pre-coding discover phase only.

danielvm-git/bigpowers · 64 tokens

slice-tasks

PLANNING SPINE STEP 2 of 3 — Slice the work: break a scoped PRD into vertical-slice stories in specs/epics/. Use after scope-work (step 1), before plan-work (step 3). Not a substitute for scope-work or plan-work.

danielvm-git/bigpowers · 59 tokens

terse-mode

Fallback ultra-compressed communication mode. Cuts token usage 75% by dropping filler, articles, and pleasantries while keeping full technical accuracy. Use ONLY when context is critically long and compressing output is necessary to continue. Not a strategy — token discipline comes from code shape (small functions…

danielvm-git/bigpowers · 101 tokens