The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill documentation directs users to download the prerequisite 'Clips Desktop' app from https://clips.agent-native.com/download. This is the official vendor-managed site for the required software.
[COMMAND_EXECUTION]: The skill provides npx commands (e.g., npx -y @agent-native/core@latest) to automate the installation of the local agent connection. These commands execute packages from the author's official @agent-native npm scope and are intended for initial environment setup.
[PROMPT_INJECTION]: The skill ingests untrusted content from the user's screen through OCR, transcripts, and frame context. This creates a surface for indirect prompt injection where content displayed on the screen could influence the agent's behavior.
Ingestion points: Data is ingested via the screen_memory_recent_context and screen_memory_search_chapters tools defined in SKILL.md.
Boundary markers: The skill instructions do not explicitly define boundary markers or 'ignore' instructions for the data returned by the tools.
Capability inventory: The skill is primarily focused on data retrieval and local configuration; it does not possess high-privilege write capabilities or direct network exfiltration tools within its primary logic.
Sanitization: There is no evidence of sanitization or filtering applied to the retrieved screen text within the skill's instructions.

rewind