pricing-strategist

Pass

Audited by Gen Agent Trust Hub on Feb 19, 2026

Risk Level: SAFE
Full Analysis
  • [DATA_EXFILTRATION] (SAFE): The skill reads FOUNDER_CONTEXT.md to gain business context. While this file contains sensitive business information, the skill lacks any network-capable commands (like curl or fetch) to transmit the data externally.
  • [REMOTE_CODE_EXECUTION] (SAFE): There are no references to external scripts, package managers, or dynamic execution functions. All logic is self-contained within the prompt instructions.
  • [PROMPT_INJECTION] (SAFE): The skill uses natural instructional language to define its task execution flow. It does not contain any markers intended to bypass safety filters or override system instructions.
  • [INDIRECT_PROMPT_INJECTION] (LOW): The skill ingests data from FOUNDER_CONTEXT.md. Evidence: 1. Ingestion points: FOUNDER_CONTEXT.md. 2. Boundary markers: Absent. 3. Capability inventory: Limited to text generation/conversation. 4. Sanitization: Absent. The risk is low as the agent only generates strategy advice based on the content.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 19, 2026, 03:54 PM