The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from multiple external sources which could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: Data is collected from Meta, TikTok, Google, and LinkedIn ad libraries via the scrapecreators-api, and landing page content is fetched via fetch_webpage or curl in Phase 6.
Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions to wrap the external content before it is processed by the LLM.
Capability inventory: The skill utilizes curl for network requests and potential command execution.
Sanitization: No explicit sanitization, filtering, or escaping of the fetched external content is defined before interpolation into the analysis phases.
[COMMAND_EXECUTION]: The skill explicitly instructs the agent to use curl as a fallback method to fetch landing page content in Phase 6.
Evidence: 'Or use curl if fetch_webpage is unavailable.' in SKILL.md. While intended for fetching web content, this provides a pathway for executing shell commands with external URLs as arguments.
[EXTERNAL_DOWNLOADS]: The skill fetches data and configuration from several external ad platform APIs and transparency centers.
Evidence: Phase 1 through 4 detail operations against Meta, TikTok, Google, and LinkedIn infrastructure. These are documented neutrally as they are core to the skill's stated purpose of ad intelligence.

competitor-ad-intelligence