The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it is designed to ingest and process untrusted data from external sources (prospect websites and third-party platforms) to generate lead scores and outreach rationales.
Ingestion points: As detailed in references/data-sources.md and references/local-prospecting.md, the skill uses Firecrawl or Browserbase to extract content from URLs found during the discovery phase.
Boundary markers: The instructions do not define specific delimiters or warnings to the agent to treat external site content as potentially untrusted data.
Capability inventory: The skill possesses the capability to write findings to CSV and markdown tables, and it executes a local CLI tool (tools/clis/github-prospects.js) to perform enrichment. It also passes data to the cold-email skill.
Sanitization: There are no instructions for the agent to sanitize, escape, or validate the content retrieved from external business websites before using it in the decision-making or output phases.
[COMMAND_EXECUTION]: The skill uses a local script to perform GitHub data enrichment, which involves processing user-provided input in a shell environment.
Evidence: references/saas-prospecting.md provides the command: node tools/clis/github-prospects.js stargazers <owner/repo> --enrich --with-company --format csv. This is a vendor-provided tool (Nairon-AI) used for its intended purpose of data retrieval.

prospecting