sales-redreach

Pass

Audited by Gen Agent Trust Hub on Jun 14, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection due to its handling of external and user-supplied data which is then used to influence agent output and stored in a persistent local file.
  • Ingestion points: The skill processes untrusted input from user queries (Step 1) and describes workflows involving Reddit post content (Step 3/4).
  • Boundary markers: There are no explicit delimiters or instructions provided to the agent to treat external content strictly as data rather than instructions.
  • Capability inventory: The skill allows the agent to write and append information to the references/learnings.md file based on interactions.
  • Sanitization: The instructions do not define any sanitization, filtering, or validation steps for content before it is processed or written to the learnings file.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 14, 2026, 12:20 AM
Security Audit — agent-trust-hub — sales-redreach