The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill gathers intelligence by scraping public web content from services including LinkedIn, Apollo, Luma, and web archives. These operations are core to the stated purpose of the skill and are performed by referenced sub-capabilities.
[COMMAND_EXECUTION]: The skill coordinates an automated pipeline by orchestrating multiple other playbooks and capabilities. Installation is handled via the standard npx goose-skills command.
[PROMPT_INJECTION]: There is a potential for indirect prompt injection as the skill ingests untrusted data from the web.
Ingestion points: External data enters the context via review-scraper, linkedin-post-research, web-archive-scraper, conference-speaker-scraper, and luma-event-attendees in Phase 1 and 3.
Boundary markers: The process includes mandatory human checkpoints for cost approval (Phase 0) and strategy selection (Phase 2) before execution begins.
Capability inventory: The skill has access to email-drafting, content-asset-creator, and potential outreach tools like cold-email-outreach and linkedin-outreach (though these are restricted to pitch-packet mode by default).
Sanitization: No specific prompt sanitization or boundary delimiters for untrusted content are described in the instructions.
[SAFE]: The skill includes explicit warnings and requires user confirmation before enabling 'live campaign mode', which prevents accidental automated outreach or unintended credit spending.

client-packet-engine