The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external sources, including business websites (via Firecrawl and Browserbase), social threads, and GitHub activity. This introduces a potential surface for indirect prompt injection if external content contains malicious instructions.
Ingestion points: Prospect websites, GitHub stargazer metadata, and public forum posts (e.g., Reddit, Hacker News).
Boundary markers: The instructions provide structured frameworks (Phase 1-5) and pass/fail checklists for qualification, which help focus the agent on data extraction rather than instruction following.
Capability inventory: The skill can execute local scripts (github-prospects.js), call various external APIs (Apollo, Truelist, etc.), and generate CSV files.
Sanitization: No explicit sanitization or filtering of external content is described; the skill relies on the agent's structured processing rules.
[COMMAND_EXECUTION]: The skill instructs the agent to use a bundled Node.js utility, tools/clis/github-prospects.js, to programmatically retrieve and enrich GitHub data. This is a controlled execution of a local tool included with the skill's infrastructure.
[EXTERNAL_DOWNLOADS]: The skill utilizes several well-known third-party services and APIs for lead enrichment and verification, including Apollo, ZoomInfo, Clay, Clearbit, and Truelist. These references are essential to the skill's function and target established, reputable service providers.

prospecting