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 instructions/trevhud/rote/agents-mdgit clone --depth 1 https://github.com/trevhud/roteWhat 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 | $0.04235 | $0.04235 |
| Opus 5 | $0.02117 | $0.02117 |
| Sonnet 5 | $0.00847 | $0.00847 |
| Haiku 4.5 | $0.00424 | $0.00424 |
Grade A, and why
rote 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 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.
How it starts
The opening of the file, as written. The whole thing — 316 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — driving rote as an installed tool
This file is for a coding agent (Codex, Cursor, a plain API agent, or
Claude Code without the plugin) that has been asked to use rote to
compile a skill. It covers the operational facts rote --help can't tell
you. If you are editing rote itself, read CLAUDE.md instead — that's the
contributor manual and does not apply here.
rote compiles an Anthropic-style skill (SKILL.md + optional
references/) into a deterministic, durable workflow: a pipeline.yaml IR
plus runnable runtime code. The north-star command is rote compile.
Invocation contract
Run it with uvx — nothing to install but uv:
uvx --from rote-cli rote <args>
- The PyPI package is
rote-cli; the command (and import name) isrote. The published wheel ships norote-cliexecutable, souvx rote-cli ...fails. Always--from rote-cli rote. - Unreleased features: swap the origin, same command —
uvx --from git+https://github.com/trevhud/rote rote <args>. - Do not clone or build a venv;
uvxhandles isolation.
Preflight first: rote doctor
Before spending money on a compilation, run the read-only preflight — it costs nothing and tells you whether a run can even succeed:
uvx --from rote-cli rote doctor --json
--json gives {version, python, drivers:[{name,available,reason}], runtimes, mcp_servers:[{name,url,auth}], apps, ok}. Gate on ok (true iff at least one
compiler driver is available — rote doctor also exits non-zero when none is).
Each unavailable driver's reason is an actionable fix. It also surfaces MCP
servers whose auth is expired/not authenticated (those would park a run) and
registered apps whose directory has gone missing. It does NOT probe CLI
subscription auth for claude/codex — that's only verified at run time.
Optional but recommended: rote baseline before compiling
uvx --from rote-cli rote baseline <skill-dir> --out <out-dir> --input task.json --json
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.
- 2d ago First seen · 316 lines · 4,235 tokens per session scan A 16f3d60be335
rote AGENTS.md is an instructions file published in the GitHub repository trevhud/rote (6 stars, last pushed 3d ago), licensed Apache-2.0. It adds 4,235 tokens to every session, about $0.0212 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.
Other instructions, from other repositories
ii-agent AGENTS.md
Instructions for Intelligent-Internet/ii-agent, covering ii-agent contributor guide, quick start, repository map, mandatory rules and architecture.
ToolRegistry AGENTS.md
Instructions for Oaklight/ToolRegistry, covering agents.md — toolregistry, what this project is, architecture, repository layout and setup and commands.
GENesis-AGI AGENTS.md
AGENTS.md instructions for WingedGuardian/GENesis-AGI, covering agent instructions, code review mandate (adversarial), gitnexus — code intelligence (advisory), genesis capability surface and skills.
gigaxity-deep-research AGENTS.md
Instructions for yoloshii/gigaxity-deep-research, covering gigaxity deep research — agent reference, tool surface, when to call which tool, environment variables and anti-patterns.
apex-accelerator instructions.instructions.md
Guidelines for creating high-quality custom instruction files for GitHub Copilot.
Framework_Development_Guide
Instructions for rezaho/MARSYS, covering 0. golden principle, 1. layered architecture (mental model), 2. core classes & responsibilities, 2.1 module overview & logical flow and src/agents/ - agent architecture & communication.