lov-expense-report

Pass

Audited by Gen Agent Trust Hub on Sep 7, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: No malicious patterns, obfuscation, or unauthorized data access were identified in the skill's scripts or instructions.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input in the form of invoice images and text descriptions.
  • Ingestion points: Invoice photos and user-provided text as described in the workflow in SKILL.md.
  • Boundary markers: The workflow includes a mandatory Step 3 where the agent must present a table of extracted data and use AskUserQuestion to obtain user confirmation before proceeding.
  • Capability inventory: The local script scripts/generate_report.py performs file system writes to create the Excel (.xlsx) output.
  • Sanitization: Data is handled through standard JSON parsing and openpyxl library methods, which prevents direct execution of input data during the report generation process.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 7, 2026, 09:40 PM
Security Audit — agent-trust-hub — lov-expense-report