ecom
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious instructions, obfuscation, or dangerous behaviors were detected in the skill files. The skill is designed for data analysis and uses local reference files for its templates.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from user-provided CSV transaction files to produce analytical reports.
- Ingestion points: The 'ecom' CLI tool reads transaction data from CSV files (e.g., 'orders.csv') to generate a 'review.json' file.
- Boundary markers: The skill relies on structured schemas ('review-schema.md') and specific report templates ('report-format.md') to constrain how the agent interprets and presents the results.
- Capability inventory: The skill uses 'Bash' to execute the analytics engine and 'Write' to generate the final markdown reports.
- Sanitization: The computational engine acts as an intermediary layer, extracting quantitative metrics before the agent processes the information, which reduces the surface for direct instruction injection from the data.
Audit Metadata