inventory-planner

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • Indirect Prompt Injection Surface: The skill processes data from external systems and user-provided files that could potentially contain malicious instructions designed to manipulate the agent's output.
  • Ingestion points: As detailed in SKILL.md and reference/data_sources.md, the skill reads sales history, SKU names, and vendor information from Shopify, Square, NetSuite, and CSV uploads.
  • Boundary markers: The instructions do not implement specific delimiters or 'ignore embedded instruction' blocks when interpreting data fields like item descriptions or names, which are then used in logic or drafting.
  • Capability inventory: The agent uses this data to draft vendor-facing emails and purchase orders, and sends metadata to an external ap-processor skill/tool.
  • Sanitization: The skill lacks explicit sanitization or validation steps to ensure that data fetched from third-party APIs or files does not contain natural language instructions that could alter the 'voice' or content of outgoing vendor emails.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 07:05 PM
Security Audit — agent-trust-hub — inventory-planner