aiskillstore/marketplace / skills/0xselenicdove/auv-computer-control
auv-computer-control
VerifiedIncorporate moeru-ai/AUV into native desktop app control through its CLI or connected MCP for observation, targeted input, reusable operations, and verification. Use for native app automation, repeated desktop workflows, or explicit AUV requests.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 6.3 KB |
| agents/openai.yaml | config | 234 B |
| references/operations.md | markdown | 4.1 KB |
| skill-report.json | config | 23.0 KB |
skill-report.json
View on GitHub{
"schema_version": "2.0",
"meta": {
"generated_at": "2026-10-06T21:36:09.592Z",
"slug": "0xselenicdove-auv-computer-control",
"source_url": "https://github.com/0xselenicdove/auv-computer-control-skill/tree/538428ee3b8d7db489b9bc12c537e25a0d86baab/skills/auv-computer-control",
"source_ref": "538428ee3b8d7db489b9bc12c537e25a0d86baab",
"model": "codex",
"analysis_version": "3.0.0",
"source_type": "community",
"content_hash": "07ebbda3ba15daa7c4de0ca63bd6d4d408cfe722408d73d470df3febb5c7aedd",
"tree_hash": "c09bdc28ba45aa24ad0e3ea72b469a04b34dc896dcd43f4b00baa33cc6c661d4",
"provenance": {
"model_effective": "codex",
"fallback_chain": [
{
"agent": "codex",
"outcome": "success"
}
],
"analysis_version": "3.0.0"
},
"upstream_version_raw": null,
"upstream_version_normalized": null,
"upstream_version_source": "none",
"upstream_version_status": "missing",
"upstream_commit_sha": "538428ee3b8d7db489b9bc12c537e25a0d86baab"
},
"skill": {
"name": "auv-computer-control",
"description": "Incorporate moeru-ai/AUV into native desktop app control through its CLI or connected MCP for observation, targeted input, reusable operations, and verification. Use for native app automation, repeated desktop workflows, or explicit AUV requests.",
"summary": "Guide authorized native desktop workflows with AUV observation, targeted input, reusable operations, and independent outcome verification.",
"icon": "π¦",
"author": "0xselenicdove",
"license": "MIT",
"category": "productivity",
"tags": [
"desktop-automation",
"auv",
"native-apps",
"computer-control",
"mcp"
],
"supported_tools": [
"claude",
"codex",
"claude-code"
],
"risk_factors": [
"external_commands",
"network"
]
},
"security_audit": {
"risk_level": "safe",
"is_blocked": false,
"safe_to_publish": true,
"analysis_status": "ok",
"summary": "All 16 static findings are false positives from documentation, Markdown delimiters, or intended desktop operations. No evidence found of credential theft, screenshot upload, prompt injection, or unauthorized reconnaissance in the reviewed passages. AUV installation, OS permissions, and stored desktop evidence still require careful user control; external runtime implementations were not audited.",
"remediation": [],
"risk_factor_evidence": [
{
"factor": "external_commands",
"evidence": [
{
"file": "references/operations.md",
"line_start": 25,
"line_end": 25
},
{
"file": "SKILL.md",
"line_start": 20,
"line_end": 26
},
{
"file": "SKILL.md",
"line_start": 26,
"line_end": 30
},
{
"file": "SKILL.md",
"line_start": 30,
"line_end": 30
},
{
"file": "SKILL.md",
"line_start": 36,
"line_end": 38
},
{
"file": "SKILL.md",
"line_start": 38,
"line_end": 39
},
{
"file": "SKILL.md",
"line_start": 39,
"line_end": 39
},
{
"file": "SKILL.md",
"line_start": 44,
"line_end": 44
},
{
"file": "SKILL.md",
"line_start": 46,
"line_end": 46
}
]
},
{
"factor": "network",
"evidence": [
{
"file": "SKILL.md",
"line_start": 3,
"line_end": 3
},
{
"file": "SKILL.md",
"line_start": 8,
"line_end": 8
}
]
}
],
"static_findings": [
{
"id": "external_commands:references/operations.md:25:powershell-invocation",
"category": "external_commands",
"severity": "high",
"pattern": "PowerShell invocation",
"file": "references/operations.md",
"line_start": 25,
"line_end": 25,
"snippet": "Linux/Windows have native drivers, but individual core commands have narrower platform support. Foll"
},
{
"id": "sensitive:references/operations.md:43:certificate-key-files",
"category": "sensitive",
"severity": "high",
"pattern": "Certificate/key files",
"file": "references/operations.md",
"line_start": 43,
"line_end": 43,
"snippet": "For ordered keyboard actions, inspect `input.keyboard --help`; `--actions` accepts tagged JSON actio"
},
{
"id": "blocker:references/operations.md:23:screen-capture-upload",
"category": "blocker",
"severity": "high",
"pattern": "Screen capture upload",
"file": "references/operations.md",
"line_start": 23,
"line_end": 23,
"snippet": "The user grants Accessibility for AX/input, Screen Recording for capture/OCR, and Automation for rel"
},
{
"id": "blocker:references/operations.md:49:system-reconnaissance",
"category": "blocker",
"severity": "low",
"pattern": "System reconnaissance",
"file": "references/operations.md",
"line_start": 49,
"line_end": 49,
"snippet": "- `auv devices list` discovers daemon-visible Devices. Root selection precedes the subcommand: `auv "
},
{
"id": "external_commands:SKILL.md:20:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 20,
"line_end": 26,
"snippet": "```sh"
},
{
"id": "external_commands:SKILL.md:26:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 26,
"line_end": 30,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:30:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 30,
"line_end": 30,
"snippet": "Inspect `auv invoke <command-id> --help` before first use in a session. Installed help is authoritat"
},
{
"id": "external_commands:SKILL.md:36:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 36,
"line_end": 38,
"snippet": "1. Establish the intended app/window from fresh observations, such as `auv invoke window.list --json"
},
{
"id": "external_commands:SKILL.md:38:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 38,
"line_end": 39,
"snippet": "3. Focus the intended control explicitly. An app/window target selects the recipient, not the text f"
},
{
"id": "external_commands:SKILL.md:39:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 39,
"line_end": 39,
"snippet": "4. Deliver the smallest targeted action. Use `app:<observed-application-id>` or `window:<observed-wi"
},
{
"id": "external_commands:SKILL.md:44:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 44,
"line_end": 44,
"snippet": "Use invoke `--json`. Retain stdout even on nonzero process exit and keep stderr separately. The insp"
},
{
"id": "external_commands:SKILL.md:46:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 46,
"line_end": 46,
"snippet": "Use task-local `--store-root` if evidence would clutter the working directory; default recording is "
},
{
"id": "network:SKILL.md:3:python-http-libraries",
"category": "network",
"severity": "low",
"pattern": "Python HTTP libraries",
"file": "SKILL.md",
"line_start": 3,
"line_end": 3,
"snippet": "description: Incorporate moeru-ai/AUV into native desktop app control through its CLI or connected M"
},
{
"id": "network:SKILL.md:8:hardcoded-url",
"category": "network",
"severity": "low",
"pattern": "Hardcoded URL",
"file": "SKILL.md",
"line_start": 8,
"line_end": 8,
"snippet": "Use [moeru-ai/AUV](https://github.com/moeru-ai/auv) as an execution layer in the existing computer-c"
},
{
"id": "blocker:SKILL.md:50:screen-capture-upload",
"category": "blocker",
"severity": "high",
"pattern": "Screen capture upload",
"file": "SKILL.md",
"line_start": 50,
"line_end": 50,
"snippet": "- Coordinate input uses the command's logical screen/window/display basis; OCR boxes can be capture "
},
{
"id": "blocker:SKILL.md:58:system-reconnaissance",
"category": "blocker",
"severity": "low",
"pattern": "System reconnaissance",
"file": "SKILL.md",
"line_start": 58,
"line_end": 58,
"snippet": "Reuse installed plugins first. When a verified repeated flow needs custom reuse, write the smallest "
}
],
"finding_verdicts": [
{
"id": "external_commands:references/operations.md:25:powershell-invocation",
"verdict": "false_positive",
"confidence": 0.98,
"reason": "The passage describes elevated PowerShell as a Windows helper setup prerequisite; it contains no PowerShell invocation or payload. Setup is restricted to authorized tasks."
},
{
"id": "sensitive:references/operations.md:43:certificate-key-files",
"verdict": "false_positive",
"confidence": 1,
"reason": "The line documents input.keyboard and tagged keyboard actions, not certificate or private-key files. It explicitly forbids submission after unverified field edits."
},
{
"id": "blocker:references/operations.md:23:screen-capture-upload",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "This passage explains user-granted screen capture and accessibility permissions, not screenshot upload. It explicitly prohibits automatic OS permission changes."
},
{
"id": "blocker:references/operations.md:49:system-reconnaissance",
"verdict": "false_positive",
"confidence": 0.98,
"reason": "Device discovery and display listing select an observed remote target for the requested workflow. The text requires preserving that selection and treats pairing as separate setup."
},
{
"id": "external_commands:SKILL.md:20:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 1,
"reason": "The backticks open a Markdown shell example containing AUV availability, version, diagnostic, help, and plugin checks. There is no Ruby execution or shell substitution."
},
{
"id": "external_commands:SKILL.md:26:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 1,
"reason": "The matched backticks close a Markdown example, followed by authorized setup guidance and CLI help documentation. They are not executable shell substitutions."
},
{
"id": "external_commands:SKILL.md:30:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 1,
"reason": "Inline backticks format AUV help and plugin discovery commands. The instructions require checking supported arguments and do not construct executable commands from untrusted input."
},
{
"id": "external_commands:SKILL.md:36:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Inline commands document window observation and explicit text-control focus. The workflow disambiguates windows and treats UI text as data, not instructions."
},
{
"id": "external_commands:SKILL.md:38:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 1,
"reason": "Backticks format focus operations and observed target identifiers. These are scoped desktop-control instructions, not shell backtick execution."
},
{
"id": "external_commands:SKILL.md:39:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 1,
"reason": "The line documents target syntax and operation-specific targeting limits. It requires observed identifiers and preserving remote Device identity, with no executable backtick expression."
},
{
"id": "external_commands:SKILL.md:44:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 1,
"reason": "Backticks format JSON result fields and status values. The passage explains result handling and independent verification, not command execution."
},
{
"id": "external_commands:SKILL.md:46:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Backticks format local recording options, storage paths, and artifact fields. The text directs local image inspection and contains no shell substitution or upload instruction."
},
{
"id": "network:SKILL.md:3:python-http-libraries",
"verdict": "false_positive",
"confidence": 1,
"reason": "The matched line is YAML prose describing AUV CLI and MCP integration. It imports no Python HTTP library and makes no network request."
},
{
"id": "network:SKILL.md:8:hardcoded-url",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "The GitHub URL identifies the upstream AUV project in a Markdown link. It is not an upload endpoint or executable request carrying user data."
},
{
"id": "blocker:SKILL.md:50:screen-capture-upload",
"verdict": "false_positive",
"confidence": 1,
"reason": "The line explains screenshot pixel conversion and logical input coordinates. It contains no upload command, destination, or instruction to transmit captures."
},
{
"id": "blocker:SKILL.md:58:system-reconnaissance",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "The passage recommends task-scoped reuse of verified workflows using argument arrays and explicit targets. It requires bounded waits and stops on ambiguous partial effects."
}
],
"semantic_findings": [],
"critical_findings": [],
"high_findings": [],
"medium_findings": [],
"low_findings": [],
"dangerous_patterns": [],
"files_scanned": 3,
"total_lines": 117,
"audit_model": "codex",
"audited_at": "2026-10-06T21:36:09.592Z",
"agent_auto_install_policy": "allowed",
"manual_install_policy": "allowed"
},
"content": {
"user_title": "Control Native Desktop Apps with AUV",
"value_statement": "Native desktop workflows can fail when automation uses stale targets or mistakes input delivery for success. This skill guides AUV observation, targeted actions, and independent verification.",
"seo_keywords": [
"Claude",
"Codex",
"Claude Code",
"AUV",
"desktop automation",
"native app control",
"OCR targeting",
"MCP integration",
"workflow verification",
"remote device control"
],
"actual_capabilities": [
"Guides discovery of installed AUV commands, plugins, platform support, and permission requirements.",
"Guides window observation, screenshot capture, OCR text discovery, and operation-specific target selection.",
"Coordinates explicit field focus, targeted text entry, keyboard actions, and grounded clicks.",
"Separates command delivery from independently verified application outcomes using results and observation.",
"Maintains observed remote Device selection and stops when that target becomes unavailable.",
"Guides task-scoped reuse through installed plugins or small scripts with bounded waits and semantic checks."
],
"limitations": [
"Installing this skill does not install AUV, app plugins, or OS permissions.",
"Command availability, targeting, and background input support vary by platform and installed version.",
"Screenshots and command output can contain sensitive data; default evidence recording uses a local .auv/store directory.",
"Successful commands and dry runs do not prove task completion; unsupported controls require another available tool or a reported blocker."
],
"use_cases": [
{
"title": "Enter Information in a Native App",
"description": "Observe the intended window, focus a field, enter approved text, and verify its complete value before submission.",
"target_user": "Office staff working with desktop applications"
},
{
"title": "Check Desktop Workflow Outcomes",
"description": "Inspect application state after targeted actions and distinguish delivered input from a verified change.",
"target_user": "QA engineers validating native application workflows"
},
{
"title": "Repeat an Authorized Remote Workflow",
"description": "Preserve an observed Device target while turning a demonstrated workflow into a bounded, task-scoped operation.",
"target_user": "IT operators managing authorized remote desktops"
}
],
"prompt_templates": [
{
"title": "Check AUV Readiness",
"prompt": "Check whether AUV is available and identify permissions needed for observing [app]. Do not install software or change permissions.",
"scenario": "Beginner: read-only runtime and permission assessment"
},
{
"title": "Fill and Verify a Field",
"prompt": "In [app], locate [window] and enter [text] into [field]. Verify the full value without submitting, and stop if the target is ambiguous.",
"scenario": "Intermediate: targeted native application input"
},
{
"title": "Recover from an Unverified Action",
"prompt": "Inspect [app] after the previous action. Check partial progress and focus before proposing a retry. Do not repeat submission or other irreversible actions.",
"scenario": "Advanced: recovery without duplicate effects"
},
{
"title": "Reuse a Verified Remote Workflow",
"prompt": "For authorized Device [device], inspect installed plugins for [verified workflow]. Propose task-scoped reuse with explicit targets, bounded waits, and independent checks.",
"scenario": "Advanced: remote target preservation and reusable automation"
}
],
"output_examples": [
{
"input": "Show an example completion report for entering a query without submitting it.",
"output": [
"Illustrative result: The intended application window was identified from fresh observation.",
"The query field was focused, and its full value was independently checked.",
"No submission was performed."
]
},
{
"input": "Show an example report when a selected remote Device becomes unavailable.",
"output": [
"Illustrative result: The selected remote Device is unavailable, so further actions stopped.",
"No fallback to the local desktop occurred.",
"The workflow remains incomplete until the selected Device is accessible."
]
}
],
"best_practices": [
"Read installed command help and refresh observations before selecting a window, control, or remote Device.",
"Verify complete field values and final application state; stop before repeating actions with uncertain effects.",
"Use authorized targets and a controlled evidence directory; protect stored screenshots and output containing sensitive information."
],
"anti_patterns": [
"Reusing stale coordinates or identifiers after changing tools, windows, display scale, or remote targets.",
"Treating successful command delivery or a dry run as proof that the user goal was achieved.",
"Changing permissions, starting listeners, enrolling credentials, or retrying submissions without appropriate authorization."
],
"faq": [
{
"question": "Does installing this skill install AUV?",
"answer": "No. AUV, its plugins, and required OS permissions are separate setup requirements."
},
{
"question": "Does the skill require MCP?",
"answer": "No. The CLI supports local core operations. An already connected MCP server can be used after discovering its actual tool schemas."
},
{
"question": "Which operating systems can it guide?",
"answer": "The references describe macOS, Linux, and Windows support. Individual commands and targeting options vary; installed help determines available behavior."
},
{
"question": "Which permissions may AUV need?",
"answer": "On macOS, operations may require Accessibility, Screen Recording, or Automation permissions. The user grants permissions; the skill forbids automatic changes."
},
{
"question": "Where is desktop evidence stored?",
"answer": "Default recording uses .auv/store under the working directory. A task-local store root can be selected; stored captures may contain sensitive data."
},
{
"question": "How does the skill confirm success?",
"answer": "It checks direct results and independently observes the application outcome. Delivered input alone is not treated as verified completion."
}
]
},
"file_structure": [
{
"name": "agents",
"type": "dir",
"path": "agents",
"children": [
{
"name": "openai.yaml",
"type": "file",
"path": "agents/openai.yaml",
"lines": 5
}
]
},
{
"name": "references",
"type": "dir",
"path": "references",
"children": [
{
"name": "operations.md",
"type": "file",
"path": "references/operations.md",
"lines": 51
}
]
},
{
"name": "SKILL.md",
"type": "file",
"path": "SKILL.md",
"lines": 61
}
]
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why β repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" β or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself β your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |