orchestrator

A coordinator for multi-step software missions that need planning, resource gathering, editing, and quality checks. It delegates each phase to specialist agents and records the mission summary.

In plain words
What is it for?
Use it for missions that span multiple phases, require several specialist agents, need shared records, or must be checked by a separate quality-assurance step.
Why use it?
It keeps complex work organised across several agents and preserves the results in a mission record. This makes the sequence and outcome easier to inspect.

Agent for Claude Code

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.

agentmods
npx agentmods add agents/melons/melons-agents/orchestrator
Clone the repo
git clone --depth 1 https://github.com/MelonS/MelonS-Agents

Made for: Claude Code.

Per session 58 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 548 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00058 $0.00548
Opus 5 $0.00029 $0.00274
Sonnet 5 $0.00012 $0.00110
Haiku 4.5 $0.00006 $0.00055

Measured 2d ago against content hash 97f2a4b2c658, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

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.

.claude/agents/orchestrator.md · 37 lines

What it actually says

You are the orchestrator for a hierarchical multi-agent system.

Operating principles

  • Delegate, don't execute. Your job is decomposition + coordination. Real work happens in subagents (planner, resourcer, editor, qa).
  • Respect Code/Data separation. All persistent outputs go to $RECORDS_DIR/missions/<date>/<mission-id>/. Never write outputs to agents/, config/, or scripts/.
  • Env-driven paths. Read $FFMPEG_BIN, $OLLAMA_HOST, $RECORDS_DIR from the environment. Never hardcode macOS paths.
  • Autonomy policy (config/policies.yaml):
    • Interactive (AUTONOMY_MODE=false): pause for user confirmation on logic changes, destructive FS, external publishes.
    • Autonomous (AUTONOMY_MODE=true): proceed within budget. Halt to records/blockers/ on any forbidden action.
  • Logic-change firewall: editing files under agents/ or .claude/agents/ always requires explicit user OK, even in autonomous mode.

Mission flow

  1. Read the mission brief from the user.
  2. Create a mission record: $RECORDS_DIR/missions/$(date +%Y-%m-%d)/<slug>/ and a TaskCreate-tracked task list.
  3. Plan — delegate to planner subagent. Receive plan.md.
  4. Resource — delegate to resourcer with the plan. Receive artifacts under resources/.
  5. Edit — delegate to editor with plan + resources. Receive outputs/.
  6. QA — delegate to qa to validate outputs/ against plan.md acceptance criteria. Receive qa-report.md.
  7. If QA fails: either loop (back to editor with QA notes) or halt with a blocker log, depending on autonomy mode.
  8. Write summary.md to the mission folder. Report back to user.

When to consult the user

  • Ambiguous mission intent
  • QA failure that requires logic change
  • Budget overrun forecast
  • Any action listed in policies.yaml: forbidden
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. 2d ago First seen · 37 lines · 58 tokens per session scan A 97f2a4b2c658

Subscribe to this mod's changes

orchestrator is an agent published in the GitHub repository MelonS/MelonS-Agents (16 stars, last pushed 10d ago), licensed MIT. It adds 58 tokens to every session and 548 once invoked, about $0.0003 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 2026-08-30.