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/tropeai/trope-cua/agents-mdgit clone --depth 1 https://github.com/tropeai/trope-cuaWrote 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/tropeai/trope-cua/agents-md)<a href="https://agentmods.dev/instructions/tropeai/trope-cua/agents-md"><img src="https://agentmods.dev/badge/instructions/tropeai/trope-cua/agents-md.svg" alt="Measured on agentmods" 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 | $0.01285 | $0.01285 |
| Opus 5 | $0.00642 | $0.00642 |
| Sonnet 5 | $0.00257 | $0.00257 |
| Haiku 4.5 | $0.00128 | $0.00128 |
Grade A, and why
trope-cua 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 6d 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 — 80 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Repository Guidelines
This repo is Trope CUA: native MCP/server drivers for background-safe UI automation, target-window capture, and the visual agent cursor overlay on Windows and macOS.
Collaboration & Git Safety
- Other agents or humans may be working in this repo at the same time. Do not run destructive git operations that could discard or rewrite their work, including
git reset --hard,git clean -fd, force checkouts/restores, orgit push --force, unless explicitly requested and coordinated. - Keep commits atomic. One commit should represent one logical change that is independently reviewable.
- Stage and commit only the files owned by the current task. Use explicit pathspecs for
git addandgit commit; do not sweep unrelated dirty files into a commit. - Before committing, inspect staged changes with
git diff --cached --statand unstage or avoid unrelated pre-staged files. - If subagents are used, give each one a narrow, explicit ownership set before it edits files, and keep concurrent ownership disjoint.
- Completed subagent work should report the commit SHA, changed files, validation commands, and any follow-up risks before the subagent is closed.
Repository Layout
src/CuaDriver.Win/: .NET Windows driver, MCP tools, UIA tree, input lanes, capture, cursor overlay, config, and recording.src/CuaDriver.Win/Cursor/: visual agent cursor overlay and motion/rendering logic.src/CuaDriver.Win/Tools/: tool definitions and tool-level route orchestration.src/CuaDriver.Win/Input/: targeted window-message and input route implementations.src/CuaDriver.Win/Browser/: browser-specific safe routes and CDP helpers.src/CuaDriver.Win/Capture/: target-window screenshot/capture code.src/CuaDriver.Win/HardCases/: child-session/AppBroadcast scaffolding and probes.tests/integration/: pytest integration checks against the built executable.native/macos/trope-cua/: Swift macOS driver, app bundle, CLI, skills, and macOS integration tests.docs/: design, safety, capture, and hard-case implementation notes.scripts/: build, install, uninstall, and test scripts.
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.
- 6d ago First seen · 80 lines · 1,285 tokens per session scan A 4a0e1c9cb603
trope-cua AGENTS.md is an instructions file published in the GitHub repository tropeai/trope-cua (29 stars, last pushed 1mo ago), licensed MIT. It adds 1,285 tokens to every session, about $0.0064 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 instructions, from other repositories
SwiftAutoGUI CLAUDE.md
Instructions for NakaokaRei/SwiftAutoGUI, covering claude.md, project overview, development commands, build and test and documentation.
apple-browsers pixels.instructions.md
Instructions for duckduckgo/apple-browsers, a project described as: DuckDuckGo iOS & macOS browsers.
tuicommander AGENTS.md
AGENTS.md instructions for sstraus/tuicommander, covering tuicommander — project rules, doc sync, tests, web-ui testing with agent-browser (browser mode, not tauri) and visual.
Shinken CLAUDE.md
Claude Code instructions for Meirtz/Shinken, covering claude.md — guide for ai coding sessions in this repo, ⛔ the one hard rule: this is a public open-source project, layout, development commands and architecture (how the pieces connect).
SwiftAutoGUI AGENTS.md
Instructions for NakaokaRei/SwiftAutoGUI: The canonical repository instructions live in CLAUDE.md.
oculos AGENTS.md
Instructions for huseyinstif/oculos, covering oculos — ai agent instructions, core workflow, discovery, list all windows and search for elements (preferred — fast).