The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external websites through firecrawl_scrape and firecrawl_map (e.g., competitor homepages, pricing pages, and review sites like G2).
Ingestion points: External URLs provided by the user or discovered via mapping are scraped into raw markdown files within the competitor-profiles/raw/ directory.
Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when the agent synthesizes the scraped content into final profile documents.
Capability inventory: The skill has the capability to write files to the local disk (synthesized profiles and raw scrapes) and perform network operations via external tools.
Sanitization: The skill lacks explicit instructions for sanitizing or filtering the scraped content before the LLM processes it for synthesis, which could allow malicious instructions embedded in a website to influence the agent's output.

competitor-profiling