pricing-pocket-waterfall

Pass

Audited by Gen Agent Trust Hub on Sep 8, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the Bash tool to execute a local Python script (scripts/pocket_price_waterfall.py) intended for data processing and analysis.
  • [INDIRECT_PROMPT_INJECTION]: The skill contains an attack surface for indirect prompt injection due to the ingestion and processing of untrusted external data.
  • Ingestion points: The skill reads external data from customers.csv, which includes user-controlled fields such as customer IDs, discount reasons, and JSON-formatted discount steps.
  • Boundary markers: There are no explicit delimiters or specific instructions for the agent to disregard potential commands or malicious formatting within the ingested CSV data.
  • Capability inventory: The skill is configured with Read, Write, and Bash capabilities, which could be abused if the agent follows instructions embedded in the data.
  • Sanitization: The skill instructions do not include steps for validating, escaping, or sanitizing the input data before it is incorporated into the cohort-level waterfall reports.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 8, 2026, 10:28 PM
Security Audit — agent-trust-hub — pricing-pocket-waterfall