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 agents/kilo-org/kilocode/orchestrator-modegit clone --depth 1 https://github.com/Kilo-Org/kilocodeWhat 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 | $0.00020 | $0.00472 |
| Opus 5 | $0.00010 | $0.00236 |
| Sonnet 5 | $0.00004 | $0.00094 |
| Haiku 4.5 | $0.00002 | $0.00047 |
Grade A, and why
orchestrator-mode 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 yesterday.
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 — 32 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Orchestrator Mode (Deprecated)
{% callout type="warning" title="Deprecated — scheduled for removal" %} Orchestrator mode is deprecated and will be removed in a future release. In the VSCode extension and CLI, agents with full tool access (Code, Plan, Debug) can now delegate to subagents automatically. You no longer need a dedicated orchestrator — just pick the agent for your task and it will coordinate subagents when helpful. (Read-only agents like Ask do not support delegation.) {% /callout %}
What Changed
Previously, orchestrator mode was the only way to break complex tasks into subtasks. You had to explicitly switch to orchestrator mode, which would then delegate work to other modes like Code or Architect.
Now, subagent support is built into agents that have full tool access (Code, Plan, Debug). When one of these agents encounters a task that would benefit from delegation — like exploring a codebase, running a parallel search, or handling a subtask in isolation — it can launch a subagent directly using the task tool. There's no need to switch agents first.
What You Should Do
- Just pick the right agent for your task. Use Code for implementation, Plan for architecture, Debug for troubleshooting. Each will orchestrate subagents where it makes sense.
- Add custom subagents if you want specialized delegation behavior. See Custom Subagents for details.
- Stop switching to orchestrator mode before complex tasks. Your current agent already has that capability.
How Subagents Work
- The agent analyzes a complex task and decides a subtask would benefit from isolation.
- It launches a subagent session using the
tasktool (e.g.,generalfor autonomous work,explorefor codebase research). - The subagent runs in its own isolated context — separate conversation history, no shared state.
- When done, the subagent returns a summary to the parent agent, which continues its work.
Agents can launch multiple subagent sessions concurrently for parallel work.
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.
- yesterday First seen · 32 lines · 20 tokens per session scan A ec0112eac53f
orchestrator-mode is an agent published in the GitHub repository Kilo-Org/kilocode (27,081 stars, last pushed 2d ago), licensed MIT. It adds 20 tokens to every session and 472 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 2026-08-30.
Other agents, from other repositories
Audit-Commits
Compares HEAD against a tag, identifies user-facing commits, ensures they have linked issues and CHANGELOG entries.
backseat-dev
Backseat Driver VS Code extension development — ClojureScript shadow-cljs, Ex action/effect framework, MCP socket server, Calva REPL integration, structural editing tools, Datascript output log, skill authoring, REPL-driven interactive programming. Use when: developing Backseat Driver features, debugging extension…
deep-review-orchestrator
Use this agent when the user asks to perform a comprehensive deep review using multiple competing AI models. Trigger phrases include: 'perform a deep review with multiple perspectives' 'get competing reviews of these changes' 'conduct a thorough review using different models' 'review these changes with multiple LLMs'…
inspector-driver
Drives the vscode-inspector MCP against a live VS Code + GitLens instance and returns raw evidence (probe values, console/log excerpts, geometry, measurements) as text. Default executor for live-inspection driving so the Opus orchestrator only reasons over distilled results. Returns evidence; does NOT judge, fix, or…
bd-tester
Agent for testing Backseat Driver tools.
bootstrap
Mission: Initialize and converge a governed, structured project context for GAAI.