The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill relies on third-party Apify actors for data extraction, including apimaestro/linkedin-posts-search-scraper-no-cookies, harvestapi/linkedin-post-search, and apimaestro/linkedin-profile-posts. These are external scripts hosted on the Apify platform that are not maintained by the skill author or a verified trusted organization.
[COMMAND_EXECUTION]: The skill triggers external tools and automated browser sessions via the --engine=chrome and --engine=apify flags to navigate LinkedIn and scrape content.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The skill reads external LinkedIn posts and top comments to generate draft responses, which could contain adversarial instructions designed to influence the AI's output (SKILL.md, Section B1/B2).
Boundary markers: Relies on instructions in engage-playbook.md and a specific 'quality gate' (anticopywriting-ai) to filter drafts, but does not implement hard delimiters or escaping for the ingested content.
Capability inventory: The agent has access to local file writing (engage-state.json), Notion database integration, and the ability to execute web scraping via Chrome and Apify.
Sanitization: Employs an AI-based review process to check for forbidden phrases (e.g., 'great post', pitch/links) and tone, but lacks programmatic sanitization of raw input data.

content-engage