competitor-analysis
Pass
Audited by Gen Agent Trust Hub on Mar 11, 2026
Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
- [NO_CODE]: The skill is composed exclusively of Markdown files that provide instructions and templates. It contains no executable scripts or system configuration changes, significantly limiting the technical attack surface.
- [PROMPT_INJECTION]: The research workflow involves fetching data from untrusted external sources (competitor websites), which creates a surface for indirect prompt injection.
- Ingestion points: External competitor URLs and search results (SKILL.md, Step 1 and 4).
- Boundary markers: Absent; the prompts do not instruct the agent to use specific delimiters for external content.
- Capability inventory: Confined to browsing and reporting; no evidence of file-write, command execution, or network exfiltration of sensitive local data.
- Sanitization: Not specified in the skill instructions.
Audit Metadata