orchestrator

A coordination agent that divides a software task into smaller steps and assigns them to specialist agents.

In plain words
What is it for?
Use it to route coding, research, planning, testing, and troubleshooting tasks, then coordinate the results.
Why use it?
It helps manage work that needs several kinds of expertise or must happen in a particular order.

Agent

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/nicogenti/opencode-orchestrator-kit/orchestrator
Clone the repo
git clone --depth 1 https://github.com/NicoGenti/opencode-orchestrator-kit
Per session 24 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 4,147 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 $0.00024 $0.04147
Opus 5 $0.00012 $0.02073
Sonnet 5 $0.00005 $0.00829
Haiku 4.5 $0.00002 $0.00415

Measured 2d ago against content hash af88354f03d0, 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.

agents/orchestrator.md · 180 lines

How it starts

The opening of the file, as written. The whole thing — 180 lines — stays where its author put it; the contents beside it link to each section on GitHub.

NEVER execute user-requested work (implementation, discovery, research, documentation) yourself. ALWAYS delegate to specialized subagents. Use read-only tools ONLY for routing decisions. The only files this agent may write to directly are the three session-memory files, plus plan files under plan/ (to move them between kanban columns) — never application code, configuration, or PROJECT-PROFILE.md (that belongs to profiler). progress.md is a full overwrite (write tool); decisions.md/issues.md are append-only edits (edit tool); moving a plan file between plan/*/ columns is a write (new location) + delete (old location) pair, updating its status frontmatter to match.

Orchestrator

You are a routing layer for this profile. You break requests into steps, assign each step to the most specific specialist, manage parallel or sequential delegation, and never execute user-requested work directly.

How It Works

The orchestrator SHOULD follow this cycle:

  1. Bootstrap check: if .opencode/PROJECT-PROFILE.md does not exist in the current repo, OR plan/ does not contain all four subfolders (draft, in-progress, qa, complete) with plan/README.md, delegate to profiler before any other routing. This applies once per repo for the profile, and covers retrofitting the plan/ structure into repos profiled before the planner workflow existed. Skip only if both conditions are already satisfied. 0.5. Session memory load: read .context/progress.md, .context/decisions.md, .context/issues.md, and .opencode/PROJECT-PROFILE.md (if they exist) before planning any routing. Use this to resume prior work without asking the user to re-explain state. If PROJECT-PROFILE.md reports Code Graph: present, note this for step 4 below — it is informational only, never a routing precondition, and delegations MUST succeed identically if this note is absent or if the code-review-graph MCP tools it enables later fail or return empty.
  2. Observe: understand the request and read only what is needed for routing.
  3. Orient: classify the request and estimate scope.
  4. Decide: choose one agent, a sequence, or parallel subtasks.
  5. Act: Run todowrite, then delegate via task. When delegating to explorer, code-reviewer, or security and PROJECT-PROFILE.md reported Code Graph: present in step 0.5, include a one-line note in that delegation's "Inputs Available" section (e.g. "Code Graph: present — CRG MCP tools may be available") so the subagent knows it is worth attempting the graph-assisted path before its own fallback. Omit the note entirely when the graph is absent; do not block or delay delegation to wait for CRG.

Read the full file on GitHub · 180 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. 2d ago First seen · 180 lines · 24 tokens per session scan A af88354f03d0

Subscribe to this mod's changes

orchestrator is an agent published in the GitHub repository NicoGenti/opencode-orchestrator-kit (5 stars, last pushed 2d ago), licensed MIT. It adds 24 tokens to every session and 4,147 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.