The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/champion_tracker.py uses importlib.util to dynamically load and execute code from a relative path (../../lead-qualification/scripts/enrich_leads.py). This dynamic loading from a computed path is a medium-risk execution pattern that bypasses static import analysis.- [DATA_EXPOSURE]: The _load_dotenv function in scripts/champion_tracker.py recursively searches up to 10 parent directories for a .env file to load environment variables. This behavior can result in the unintended loading and exposure of sensitive credentials from parent directory environments not explicitly associated with the skill.- [EXTERNAL_DOWNLOADS]: The skill performs network requests to api.apify.com to fetch LinkedIn profile data. While this is the intended functionality, it introduces a dependency on an external service and the security of its API.- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted data from external LinkedIn profiles via the Apify API.
Ingestion points: scripts/champion_tracker.py (via LinkedInEnricher.enrich_batch).
Boundary markers: None identified in script logic or prompt templates.
Capability inventory: File system read/write (snapshots/, output/, input/), Network requests (api.apify.com).
Sanitization: Relies on standard JSON and CSV libraries for structural output; no content-level validation or sanitization of profile fields is performed.

champion-tracker