competitive-analysis
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill utilizes WebFetch and WebSearch tools to perform competitor research. This creates an attack surface where an attacker could place malicious instructions on a website being analyzed to influence the agent's behavior.
- Ingestion points: WebFetch and WebSearch are used in SKILL.md and references/competitive-analysis-guide.md to gather competitor data.
- Boundary markers: No explicit sanitization or boundary markers are defined for the fetched content.
- Capability inventory: Access to Read, Glob, Grep, WebFetch, and WebSearch tools.
- Sanitization: Absent; the skill does not instruct the agent on how to escape or filter external content.
- [COMMAND_EXECUTION]: The skill provides templates for using the GitHub CLI (gh) to retrieve repository statistics, releases, and contributor counts.
- Evidence: Multiple files (SKILL.md, references/competitive-analysis-guide.md, rules/market-competitive.md) contain bash snippets for gh api and gh release commands.
- Risk: While these are standard research tools, they rely on shell execution. If the agent populates the owner or repository parameters with unsanitized untrusted input, it could potentially lead to command injection.
Audit Metadata