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 skills add appergb/desktop-agent-ops --skill skillgit clone --depth 1 https://github.com/appergb/desktop-agent-opsWrote 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/skills/appergb/desktop-agent-ops/skill)<a href="https://agentmods.dev/skills/appergb/desktop-agent-ops/skill"><img src="https://agentmods.dev/badge/skills/appergb/desktop-agent-ops/skill/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/skills/appergb/desktop-agent-ops/skill"><img src="https://agentmods.dev/badge/skills/appergb/desktop-agent-ops/skill.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 | $0.00038 | $0.05098 |
| Opus 5.5 | $0.00015 | $0.02039 |
| Sonnet 5.5 | $0.00008 | $0.01020 |
| Haiku 4.5 | $0.00004 | $0.00510 |
Grade A, and why
desktop-agent-ops 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.
This is a copy
100% identical to desktop-agent-ops — 0 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 426 lines — stays where its author put it; the contents beside it link to each section on GitHub.
This is the detailed reference manual. For the quick operations guide, see
desktop-agent-ops.md.
Desktop Agent Ops — Detailed Reference Manual
Table of Contents
- Tool Priority
- Auto-setup Gate
- Core Execution Loop
- Smart Targeting with Four-Layer Fallback
- Failure Recovery
- Generalization: How to Apply This to ANY App
- Text Input and Send Rules
- DPI / HiDPI / Retina
- CLI Reference
- Workflow Examples
- Reference Documents
- Scope
- Hard Rules
- Custom Workflows
Tool Priority
Use this skill ONLY as Priority 3 — after MCP servers/structured APIs (Priority 1) and native CLI/AppleScript (Priority 2). See desktop-agent-ops.md section 1 for the full decision framework.
Rule: Never use screen OCR to do what a structured API can do.
Auto-setup Gate
Run first_run_setup.py --check at session start. If not ready, run first_run_setup.py to auto-install all dependencies. Then set $PY. See desktop-agent-ops.md section 2 for details.
Core Execution Loop
FOCUS → LOCATE (accessibility/OCR) → BOUNDS-CHECK → MOVE → READBACK → EXECUTE → VERIFY
CRITICAL RULE: Click coordinates MUST come from accessibility or OCR output — NEVER from visual estimation of screenshots. Models frequently confuse left/right and misjudge pixel distances. Structured accessibility and OCR output return exact pixel coordinates.
CRITICAL RULE: Move → Readback → Click. Before every click, move the cursor first, read back mouse-position, verify the offset is ≤ 5px, then click. Never click without readback.
CRITICAL RULE: Re-locate before every click in multi-step tasks. Window positions, dialog states, and UI layouts change between steps. Never reuse coordinates from a previous step.
What ships with it
60 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- agents/openai.yaml 528 B
- desktop-agent-ops.md 17 KB
- references/app-names.md 9.4 KB
- references/app-wechat-desktop.md 3.3 KB
- references/app-wechat-macos.md 158 B
- references/app-wechat-windows.md 160 B
- references/chat-app-macos.md 3.5 KB
- references/cleanup-rules.md 1.2 KB
- references/collaboration-rules.md 1.4 KB
- references/coordinate-reconstruction.md 3.0 KB
- references/custom-workflows.md 4.9 KB
- references/eval-scenarios.md 2.7 KB
- references/example-cases.md 18 KB
- references/market-precision-targeting-gap-analysis.md 21 KB
- references/operation-patterns.md 5.5 KB
- references/platform-linux.md 2.0 KB
- references/platform-macos.md 3.9 KB
- references/platform-windows.md 4.2 KB
- references/precise-targeting.md 4.0 KB
- references/reproducible-setup.md 3.2 KB
- references/target-providers.md 4.7 KB
- references/validation-patterns.md 3.4 KB
- references/workflow.md 4.0 KB
- scripts/accessibility_provider.py 3.0 KB runs code
- scripts/ax_provider.py 7.6 KB runs code
- scripts/cleanup_task.py 1.0 KB runs code
- scripts/click_and_verify.py 4.6 KB runs code
- scripts/desktop_ops.py 15 KB runs code
- scripts/dispatch_agent.py 21 KB runs code
- scripts/doctor.py 5.5 KB runs code
- scripts/first_run_setup.py 25 KB runs code
- scripts/input_runtime.py 1.4 KB runs code
- scripts/linux_atspi_provider.py 5.5 KB runs code
- scripts/local_agent.py 26 KB runs code
- scripts/ocr_text.py 14 KB runs code
- scripts/permission_bootstrap.py 9.1 KB runs code
- scripts/platform_probe.py 2.8 KB runs code
- scripts/pointer_runtime.py 5.0 KB runs code
- scripts/region_diff.py 3.2 KB runs code
- scripts/resolve_python.py 1.7 KB runs code
- scripts/runtime_support.py 448 B runs code
- scripts/screen_runtime.py 5.4 KB runs code
- scripts/secret_scanner.py 7.7 KB runs code
- scripts/smoke_test.py 3.2 KB runs code
- scripts/target_provider_chain.py 8.4 KB runs code
- scripts/target_report.py 2.1 KB runs code
- scripts/target_resolver.py 3.3 KB runs code
- scripts/target_runtime.py 3.5 KB runs code
- scripts/targeting.py 3.2 KB runs code
- scripts/task_context.py 5.4 KB runs code
- scripts/task_paths.py 972 B runs code
- scripts/template_match.py 4.3 KB runs code
- scripts/text_runtime.py 7.4 KB runs code
- scripts/vision_ocr.py 8.8 KB runs code
- scripts/window_backends.py 14 KB runs code
- scripts/window_kernel.py 2.0 KB runs code
- scripts/window_regions.py 2.3 KB runs code
- scripts/windows_uia_provider.py 7.5 KB runs code
- scripts/workflow_executor.py 4.7 KB runs code
- scripts/workflow_loader.py 13 KB runs code
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 · 426 lines · 38 tokens per session scan A ff6cbabbccd4
desktop-agent-ops is a skill published in the GitHub repository appergb/desktop-agent-ops (17 stars, last pushed 6mo ago), licensed MIT. It adds 38 tokens to every session and 5,098 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 0 findings. It is 100% identical to desktop-agent-ops, differing in 0 lines, and is treated as a copy.
Other skills, from other repositories
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
next-cache-components-optimizer
Optimize the meaningful UI available immediately from a Next.js route on an initial load (hard navigation) or named client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route and entry point at a time; the shipped test then guards…