competitor-profiling
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from external competitor websites and third-party review platforms, creating a potential vector for malicious instructions to influence agent behavior.
- Ingestion points: Content is retrieved from user-provided competitor URLs and review sites (G2, Capterra, etc.) using the Firecrawl MCP tools (SKILL.md, Phase 1).
- Boundary markers: While the skill uses structured markdown templates for its final report, it does not explicitly instruct the agent to use delimiters or sanitization when processing raw scraped markdown (SKILL.md, Research Process).
- Capability inventory: The agent has the capability to write numerous files to the local filesystem (under
competitor-profiles/) and perform network queries via specialized MCP tools (SKILL.md, Saving Raw Data). - Sanitization: No specific content filtering or instruction-stripping logic is defined for the scraped site content.
Audit Metadata