margin-monitoring

Pass

Audited by Gen Agent Trust Hub on Sep 25, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a data ingestion surface for processing SKU margins, which is managed via explicit security boundaries.
  • Ingestion points: External data enters via CSV exports, manual pastes, and MCP connections specified in the 'How to run' section of SKILL.md.
  • Boundary markers: The skill requires adherence to references/agent-security.md, which provides strict instructions to treat fetched content as data rather than instructions and to use delimiters.
  • Capability inventory: The skill writes to .agents/store-loop-ledger.md and reads from .agents/product-context.md. It lacks capabilities for unauthorized network transmission or arbitrary command execution.
  • Sanitization: The referenced security policy mandates HTML-escaping for all interpolated values and URL validation to https: or relative paths.
  • [SAFE]: The skill exhibits no signs of obfuscation, remote code execution, or privilege escalation. It adheres to best practices for data persistence and secret management as outlined in its own reference files.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 25, 2026, 04:06 PM