The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructions include installing the deepline CLI tool using a custom NPM registry (https://code.deepline.com/api/v2/npm/). This registry is an external source outside of official public package repositories.- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources, specifically Google Sheets and Drive comments, to guide agent behavior and Play revisions.\n
Ingestion points: Data is ingested via the google_workspace_request tool using sheets.spreadsheets.values.get and drive.comments.list methods as described in SKILL.md.\n
Boundary markers: The instructions do not specify any delimiters or safety warnings for the agent when interpreting the content retrieved from the spreadsheet.\n
Capability inventory: The skill possesses the capability to execute tools (deepline tools execute), interact with the file system (mkdir, tee), and execute local scripts via node.\n
Sanitization: There is no explicit sanitization or validation logic described for the feedback data read from the Workspace environment before it is used to modify Play behavior.- [COMMAND_EXECUTION]: The skill uses shell commands and local script execution to facilitate its workflow. It executes node to run a local presentation script (scripts/review-sheet-presentation.mjs) and utilizes common utilities like jq, mkdir, and tee for data handling and environment setup.

deepline-plays-review