Conductor is a command-line tool for defining and running workflows in which multiple AI agents perform connected tasks. Developers use YAML files to create repeatable pipelines for activities such as code review, research and synthesis, or planning and implementation, with routing handled by explicit conditions. The catalogue add-ons provide instructions, skills, plugins, and an agent for working with Conductor.
Borrowing it
Nothing to install: this file belongs to microsoft/conductor. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/microsoft/conductor/main/AGENTS.mdgit clone --depth 1 https://github.com/microsoft/conductorWrote 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/instructions/microsoft/conductor/agents-md)<a href="https://agentmods.dev/instructions/microsoft/conductor/agents-md"><img src="https://agentmods.dev/badge/instructions/microsoft/conductor/agents-md/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/instructions/microsoft/conductor/agents-md"><img src="https://agentmods.dev/badge/instructions/microsoft/conductor/agents-md.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 | $0.01623 | $0.01623 |
| Opus 5.5 | $0.00649 | $0.00649 |
| Sonnet 5.5 | $0.00325 | $0.00325 |
| Haiku 4.5 | $0.00162 | $0.00162 |
Grade A, and why
conductor AGENTS.md 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.
How it starts
The opening of the file, as written. The whole thing — 115 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Conductor is a Python 3.12+ CLI that runs multi-agent workflows defined in YAML (parallel and for-each groups, conditional routing, loop-backs, human gates) on several LLM providers. This file is the always-loaded entry point for agents working on Conductor. It holds the rules that apply to every change and routes you to the guide that owns your task.
To author or run a workflow, use the bundled Conductor skill
(plugins/conductor/skills/conductor/SKILL.md) and the user docs in docs/.
Before you change anything
- Read every guide the task router lists for the paths you will touch. They hold the invariants, change checklists and regression tests for that area; this file does not repeat them.
- If a required guide cannot be read (no file access, missing file), say so and ask for its content. Do not guess a contract.
Universal rules
- Enforce at run time, not only in
conductor validate.conductor runnever calls the static validator (src/conductor/config/validator.py), so a new restriction must also be checked on the execution path. - Capabilities are contracts. A provider's
ProviderCapabilitiesmust describe what it actually does; if a claim cannot always be honored, declare the weaker value. A behavior or contract change to one provider applies to every provider where it is meaningful. - Keep
runandresumein parity. A flag added torunis added toresumeunless it cannot apply to a resumed run, and both paths wire the same emitters, logs and dashboard lifecycle. - Never silently widen access. Tool allowlists, MCP servers, credentials, settings tiers and filesystem grants are honored exactly; refuse a configuration you cannot honor rather than granting more than was declared.
- Surface failures. Do not swallow exceptions or report a problem the user must act on only at debug level. Optional diagnostics (records, logs, labels) must never abort a run, but their failure must be visible.
- Rich output is markup-safe. Never interpolate a runtime value into a
string Rich parses as markup. Build consoles with
conductor.console.make_console(), style withstyled(), and passTextto panel titles and prompts. Rules: console output. - Tests are hermetic. Never touch the developer's real home directory,
temp directory, running workflows, network or CLI credentials. The autouse
fixtures in
tests/conftest.pyisolate these;real_apiandinstall_scriptstests run only when selected with-m. - Style. Ruff (line length 100), Google-style docstrings, type hints
checked by
ty, Pydantic v2, async/await for provider operations. Comment only to explain why.
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.
- 4d ago Changed · -580 lines · -45,172 tokens per session scan B → A 65883b091996
- 29d ago Changed · +10 lines · +386 tokens per session 8ce6c6668424
- 1mo ago Changed · +2 lines · +192 tokens per session 6d3d0886f9e7
- 1mo ago Changed · +4 lines · +1,922 tokens per session 68a206e787eb
- 1mo ago Changed · +3 lines · +390 tokens per session d1a9d1c3bf95
- 1mo ago Changed · +6 lines · +3,044 tokens per session a18c1646768a
- 1mo ago First seen · 670 lines · 40,861 tokens per session scan B 6393704cc81d
conductor AGENTS.md is an instructions file published in the GitHub repository microsoft/conductor (476 stars, last pushed 3d ago), licensed MIT. It adds 1,623 tokens to every session, about $0.0065 per session on Opus 5.5. 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 instructions, from other repositories
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, cache eligibility and rollout and applying the notice (cutover).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode copilot-instructions.md
Copilot instructions for microsoft/vscode, covering vs code copilot instructions, project overview, root folders, core architecture (src/ folder) and built-in extensions (extensions/ folder).
deepseek-harness AGENTS.md
AGENTS.md instructions for deepseek-ai/deepseek-harness, covering agents.md, pre-stable apis and released session data, repository layout, commands and host sandbox failures.