The Agent Skills Directory

[SAFE]: The skill performs legitimate network operations to the well-known Apify service (api.apify.com) to retrieve LinkedIn profile information, which is consistent with its stated purpose of lead tracking.
[SAFE]: Sensitive information is handled securely using environment variables (APIFY_API_TOKEN). The skill includes logic to load these from a .env file and demonstrates evidence of redacting tokens in error logs within the baseline.json file.
[SAFE]: File system operations are restricted to the skill's specific directories (input/, output/, snapshots/), used for managing CSV inputs, generated reports, and historical data snapshots.
[SAFE]: Dynamic code execution is limited to importing a local sibling script (enrich_leads.py) for modularity and code reuse within the agent environment, without executing remote or untrusted code.
[INDIRECT_PROMPT_INJECTION]: The skill ingests external data from LinkedIn profiles (such as 'about' sections and headlines) and processes user-provided CSV files. While this creates a potential surface for indirect prompt injection if the agent interprets these fields as instructions, it is a standard operational requirement for enrichment tools and currently lacks high-risk execution sinks for such data.

champion-tracker