venture-sensei

Pass

Audited by Gen Agent Trust Hub on Jul 12, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests untrusted data from the internet via WebSearch and WebFetch tools (Step 1: Research). This data is then used to construct a startup critique (Step 2: Output). There are no delimiters or instructions to ignore embedded commands within the retrieved data, which could allow malicious web content to influence the agent's behavior.
  • Ingestion points: SKILL.md (Step 1: Research)
  • Boundary markers: Absent for untrusted external data.
  • Capability inventory: Tool access to WebSearch and WebFetch for data retrieval.
  • Sanitization: No sanitization or filtering of external input is performed.
  • [EXTERNAL_DOWNLOADS]: The skill utilizes network-enabled tools to download and process content from external third-party domains during the business evaluation process. This is a core part of its intended functionality for competitor and market research.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 12, 2026, 12:36 AM
Security Audit — agent-trust-hub — venture-sensei