The Agent Skills Directory

[SAFE]: The skill operates as intended for competitive intelligence. It leverages legitimate external services through defined MCP tools and maintains data locally without unauthorized exfiltration.
[PROMPT_INJECTION]: The skill possesses a standard vulnerability surface for indirect prompt injection due to its core functionality of processing external website content.
Ingestion points: Untrusted data is pulled from arbitrary competitor URLs via the firecrawl_scrape tool as described in Phase 1 of the SKILL.md file.
Boundary markers: The synthesis instructions in SKILL.md do not include specific delimiters or warnings to the model to ignore potential instructions embedded in the scraped text.
Capability inventory: The agent is instructed to write files to the local competitor-profiles/ directory, which is a necessary capability for the skill's stated purpose.
Sanitization: There are no instructions for sanitizing or filtering the scraped content before it is processed by the model to generate the final reports.

competitor-profiling