renewal-predictor

Pass

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process a wide variety of untrusted data sources from the local environment, which presents a surface for indirect prompt injection.
  • Ingestion points: The skill actively searches for and parses CSV, JSON, YAML, XLSX files, CRM exports, meeting notes, and communication logs in the working directory (SKILL.md, Execution Protocol steps 1 and 2).
  • Boundary markers: The instructions lack explicit data delimiters or 'ignore-instructions' guards for the content being processed, relying instead on the agent's ability to normalize data to a rubric.
  • Capability inventory: The skill possesses 'Bash' and 'Write' permissions, enabling it to execute system commands and create new files (renewal-forecast.md) based on the interpretation of potentially malicious input data.
  • Sanitization: No specific sanitization, escaping, or validation logic is defined for the external text content (e.g., meeting notes or summaries) before it is analyzed by the model.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 28, 2026, 07:25 AM
Security Audit — agent-trust-hub — renewal-predictor