The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the installation of software from external sources including GitHub repositories (github.com/gaboe/homebrew-acta, github.com/ggml-org/whisper.cpp) and the Chrome Web Store.
[COMMAND_EXECUTION]: The skill executes several shell commands using package managers (brew, scoop) and the specialized acta CLI to perform setup, license management, and meeting processing.
[PERSISTENCE]: The command acta serve --install is used to establish a background service or logon task, which is a documented persistence mechanism intended for the application's core functionality.
[INDIRECT_PROMPT_INJECTION]: The skill reads and processes meeting transcripts (untrusted data) using the acta show command. This represents an attack surface where instructions embedded in the meeting audio could be misinterpreted by the agent.
Ingestion points: Meeting transcripts are read from the local file system (e.g., ~/acta/<run>/transcript.md) via the acta show tool.
Boundary markers: No explicit delimiters or boundary markers are used when interpolating transcript content into the prompt.
Capability inventory: The skill can execute shell commands (brew, scoop, acta), read files, and potentially write files (e.g., acta license add).
Sanitization: No programmatic sanitization is mentioned; however, the skill documentation mandates a human-in-the-loop review step where the user must approve any tasks generated from the transcript before they are performed.

acta-diurna