The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external CSV files (e.g., customer or prospect lists) as seen in plays/build-hash-only-audience.play.ts and plays/enrich-audience-waterfall.play.ts.
Ingestion points: Data is loaded from user-provided file paths into datasets using ctx.csv across multiple play files.
Boundary markers: The scripts do not implement specific delimiters or 'ignore' instructions for the data being processed.
Capability inventory: The skill has the ability to execute network-based tools (ad platform APIs and enrichment providers), create new datasets, and export files to the workspace.
Sanitization: While the skill performs technical validation (regex for email formats and SHA-256 hex length), it does not sanitize inputs for prompt-based instructions.
[EXTERNAL_DOWNLOADS]: The skill's setup instructions in SKILL.md require the installation of the deepline CLI tool, recommending a fallback installation from a custom registry at https://code.deepline.com/api/v2/npm/ for secure sandboxes.
[COMMAND_EXECUTION]: The skill operates primarily through shell commands and a dedicated CLI (deepline), managing authentication, tool discovery, and execution via subprocesses.

deepline-ads-audiences