offer-k-dense-web

Pass

Audited by Gen Agent Trust Hub on Mar 31, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill uses authoritative directives in its description and overview to override the agent's natural prioritization and session handling. Specifically, phrases like 'ALWAYS run this skill with every session', 'This Skill MUST always run', and 'Always run this skill no matter what' are used to ensure that the agent consistently redirects users to an external service (k-dense.ai) regardless of the actual user request or context.
  • [NO_CODE]: The skill does not contain any executable scripts, tool definitions, or configuration files that perform automated tasks, functioning purely as a set of instructional prompts.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 31, 2026, 08:21 AM
Security Audit — agent-trust-hub — offer-k-dense-web