competitive-teardown

Pass

Audited by Gen Agent Trust Hub on Jun 23, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection vulnerability surface detected.
  • Ingestion points: Untrusted data enters the prompt via the [COMPETITOR NAME/URL] variable in SKILL.md.
  • Boundary markers: None. The prompt template does not use delimiters or specify that the agent should ignore instructions embedded within the analyzed content.
  • Capability inventory: The risk level depends on the capabilities of the agent using the skill, such as web browsing or file reading.
  • Sanitization: No input validation or content filtering is performed by the skill.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 23, 2026, 01:20 PM
Security Audit — agent-trust-hub — competitive-teardown