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/evilfreelancer/screenshotting-gui/agents-mdgit clone --depth 1 https://github.com/EvilFreelancer/screenshotting-guiWrote 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/evilfreelancer/screenshotting-gui/agents-md)<a href="https://agentmods.dev/instructions/evilfreelancer/screenshotting-gui/agents-md"><img src="https://agentmods.dev/badge/instructions/evilfreelancer/screenshotting-gui/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 | $0.00907 | $0.00907 |
| Opus 5 | $0.00453 | $0.00453 |
| Sonnet 5 | $0.00181 | $0.00181 |
| Haiku 4.5 | $0.00091 | $0.00091 |
Grade A, and why
screenshotting-gui 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 — 71 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Authoring & maintaining screenshotting-gui
Agent guide for editing this skill's own repo. For the catalog-level sync
process see AGENTS.md in the rpa-skills
aggregator.
What this skill is
A deterministic, general-purpose pipeline to screenshot a running GUI app on an
isolated Xvfb display, drive it with python-Xlib (XTEST), and annotate with
ImageMagick. The canonical instructions are SKILL.md (loaded by progressive
disclosure). Keep it app-agnostic - no per-application coordinate tables baked in.
Non-negotiable invariant
Target display :99 (Xvfb), never :0. The scripts default their own target to
:99; the launched app reads $DISPLAY, so the workflow always exports
DISPLAY=:99. Any change that could send input or a capture to :0 is a defect.
Layout
screenshotting-gui/
├─ SKILL.md # frontmatter (name, version, description) + workflow + gotchas
├─ scripts/
│ ├─ xctl.py # XTEST driver (move/click/key/type/scroll/drag/place/shot/winlist)
│ ├─ xpaste.py # CLIPBOARD owner for unicode paste
│ └─ annotate.py # ImageMagick: box/crop/vstack/arrow
├─ README.md # human overview
├─ .claude-plugin/ .codex-plugin/ .cursor-plugin/ # per-agent manifests (mirror version + description)
└─ LICENSE
Dependency policy
Scripts use stdlib + python-Xlib + ImageMagick only. The optional faster capture
path may use Pillow, but the default (import) must keep working without it. No
xdotool/wmctrl/pyautogui, no root, no install steps at runtime.
Testing changes
Validate on a real Xvfb, not by eye. Bring up Xvfb :99, launch a controllable GUI
(a small Tk window is enough; xterm/xclock if present), and exercise every
xctl.py command, the xpaste.py unicode path, and each annotate.py subcommand,
checking the output PNGs. A safe way to run the whole matrix is to dispatch it to a
subagent that is told the :99-only rule. Keep test artifacts out of the repo.
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 First seen · 71 lines · 907 tokens per session scan A b6831feda11f
screenshotting-gui AGENTS.md is an instructions file published in the GitHub repository EvilFreelancer/screenshotting-gui (5 stars, last pushed 22d ago), licensed MIT. It adds 907 tokens to every session, about $0.0045 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
nanobrowser CLAUDE.md
Instructions for nanobrowser/nanobrowser, covering claude.md, project overview, development commands, workspace tips and architecture.
agentic-playwright data-strategy.instructions.md
Instructions for idavidov13/agentic-playwright, covering data strategy, critical, file locations, instructions and phase 1: classify the value you need.
agentic-playwright config.instructions.md
Instructions for idavidov13/agentic-playwright, covering configuration, critical, file locations, instructions and phase 1: understand how env files are loaded.
agentic-playwright enums.instructions.md
Instructions for idavidov13/agentic-playwright, covering enums, critical, file locations, instructions and phase 1: decide if the value belongs in an enum.
inspecto CLAUDE.md
Instructions for inspecto-dev/inspecto, covering inspecto — claude code development guide, project overview, monorepo structure, development phases (load each file as needed) and key architectural decisions.
compare-mcp CLAUDE.md
Instructions for Cristophereasygoing927/compare-mcp, covering compare-mcp, architecture, running and key decisions.