competitor-profiling
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes data from external, untrusted websites through the Firecrawl and DataForSEO tools. This represents an indirect prompt injection surface where a competitor's website content could attempt to override the agent's instructions or bias the research results.\n
- Ingestion points: Content is retrieved from arbitrary competitor URLs and third-party review sites using
firecrawl_map,firecrawl_scrape, andfirecrawl_searchtools as defined inSKILL.md.\n - Boundary markers: The instructions do not specify explicit delimiters or "ignore embedded instructions" warnings to separate untrusted scraped content from the agent's primary instructions during the synthesis phase.\n
- Capability inventory: The skill has the ability to write files to the project's local file system under the
competitor-profiles/directory, as specified in the directory layout section ofSKILL.md. It does not possess code execution, shell access, or privileged system capabilities.\n - Sanitization: While the skill persists raw data to disk for auditing, it lacks explicit sanitization, escaping, or filtering steps for the ingested external content before it is interpolated into prompts.
Audit Metadata