The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill configuration initiates the download and installation of a Go-based command-line tool from a third-party GitHub repository (github.com/ossianhempel/things3-cli/cmd/things@latest).
[PRIVILEGE_ESCALATION]: The instructions advise the user to grant 'Full Disk Access' to the AI agent or Terminal. This macOS permission bypasses standard Transparency, Consent, and Control (TCC) protections, providing the software with unrestricted access to sensitive user data including mail, messages, browser history, and backups.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates an indirect prompt injection surface by ingesting data from the local Things 3 database.
Ingestion points: The agent reads task titles, notes, and metadata from the local SQLite database (ThingsData-*).
Boundary markers: There are no instructions for the agent to use delimiters or ignore embedded instructions when reading task content.
Capability inventory: The skill executes the things CLI for both read and write operations, including adding and updating tasks via URL schemes.
Sanitization: There is no evidence of filtering or sanitization of the data retrieved from the database before it is processed by the agent.
[COMMAND_EXECUTION]: The skill relies on executing the things CLI tool to perform all its primary functions, including searching, listing, and updating task data.

things-mac