Agent Orchestrator is a local desktop workspace for running and supervising multiple coding agents, giving each task its own session, workspace, branch, and feedback loop. It is for developers coordinating agent fleets across repositories, pull requests, CI runs, reviews, and merges. The catalogue entries are commands, skills, and instructions for working with this orchestrator.
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.
git clone --depth 1 https://github.com/Untrivial-ai/agent-orchestratorWrote 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/commands/untrivial-ai/agent-orchestrator/preview)<a href="https://agentmods.dev/commands/untrivial-ai/agent-orchestrator/preview"><img src="https://agentmods.dev/badge/commands/untrivial-ai/agent-orchestrator/preview/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.
<a href="https://agentmods.dev/commands/untrivial-ai/agent-orchestrator/preview"><img src="https://agentmods.dev/badge/commands/untrivial-ai/agent-orchestrator/preview.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00000 | $0.01194 |
| Opus 5.5 | $0.00000 | $0.00478 |
| Sonnet 5.5 | $0.00000 | $0.00239 |
| Haiku 4.5 | $0.00000 | $0.00119 |
Grade A, and why
preview 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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 179 lines — stays where its author put it; the contents beside it link to each section on GitHub.
ao preview
Open a URL or workspace file in the desktop browser panel for the current
session, or start a deterministic session-owned dev server from
an existing .ao/launch.json.
Static HTML and Markdown do not need a development server. Open them directly
with ao preview <workspace-path>. Never create or modify package.json,
install dependencies, or introduce npm or another server solely to display
static files.
Workspace file paths
Treat every relative file target as relative to the session workspace root, not the shell's current directory. The same command works from the workspace root or any subdirectory:
ao preview README.md
ao preview docs/guide.md
Do not compensate for the current directory with ../README.md. Relative
targets already start at the workspace root, so parent traversal instead tries
to leave the workspace and is rejected with PREVIEW_FILE_OUTSIDE_WORKSPACE.
An absolute path is also accepted when it resolves inside the session
workspace, but a workspace-root-relative path is clearer and more portable.
AO serves workspace files through the daemon's existing confined loopback
preview origin. This is a local preview, not a deployment or a new development
server. Do not open workspace files with file://; direct file URLs are not
the AO Browser panel's static-file workflow.
Use a managed server only when the project genuinely has a runtime. Start an
existing .ao/launch.json configuration when present. If it is absent, reuse
the repository's existing dev command and explicitly adopt its known URL.
Do not create .ao/launch.json unless the user asks for reusable launch
configuration.
Automatic artifact handoff
When a browser-displayable file is itself the artifact the user requested, open it immediately after creating or materially updating it:
ao preview docs/plan.md
ao preview report.html
ao preview output.pdf
ao preview diagram.svg
ao preview mockup.png
Do this without waiting for a separate "open it" request. Browser-displayable artifacts include Markdown, HTML, PDF, SVG, and common image formats such as PNG, JPEG, GIF, WebP, and AVIF. If the task produces several files, open the primary requested artifact rather than cycling through every output.
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.
- 1mo ago First seen · 179 lines · 0 tokens per session scan A b62cfb5f6be8
preview is a command published in the GitHub repository Untrivial-ai/agent-orchestrator (13,060 stars, last pushed 5d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 1,194 tokens. 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 commands, from other repositories
analyze
Analyze a webpage's structure using Actionbook's verified selectors and show available elements for scraping.
live-canvas
Conduct design interviews, generate UI variations, and collect live click-to-annotate feedback that streams into the session so edits land without leaving the browser. Use when the user wants rapid iterative UI refinement, not just batched feedback.
sync-all
Sync skills, commands, MCP servers, and preferences between Claude, Codex, Copilot, and Cursor in one operation.
sync-to-copilot
Sync skills and instructions from Claude or Codex to GitHub Copilot CLI.
sync-from-copilot
Sync skills and instructions from GitHub Copilot CLI to Claude, Codex, or Cursor.
run-e2e
Run end-to-end tests against the application using Playwright or Cypress.