swedish-sie-import-export

Pass

Audited by Gen Agent Trust Hub on Nov 18, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill provides deterministic rules for parsing and validating Swedish Standard Import Export (SIE4) files. Because these files are plain-text and contain user-controlled fields such as comments (#PROSA) and transaction descriptions, they present an indirect prompt injection surface where malicious instructions could be embedded to influence the agent's logic.\n
  • Ingestion points: External SIE4 (.SE and .SI) accounting files processed by the agent.\n
  • Boundary markers: The skill instructions define structural boundaries using record labels and quoting rules but do not include explicit instructions for the agent to disregard natural language commands within text fields.\n
  • Capability inventory: An agent using this skill typically has the capability to read and process filesystem data, which is then used to perform accounting validation.\n
  • Sanitization: While the skill includes exhaustive rules for structural and mathematical validation (e.g., verifying balance sheet continuity and verification sums), it does not prescribe sanitization for the semantic content of free-text fields.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 18, 2026, 09:59 PM