automated-reconciliation
Pass
Audited by Gen Agent Trust Hub on May 9, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill's design and instructions align with legitimate financial data processing workflows and professional accounting practices.
- [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because it ingests untrusted financial data (Source A and Source B) which is then processed by an LLM to resolve ambiguous matches. This could allow maliciously crafted data entries to influence the agent's logic.
- Ingestion points: External financial data sources such as Bank Statements (PDF/CSV) and General Ledger exports as defined in SKILL.md.
- Boundary markers: None; the skill does not instruct the agent to use specific delimiters or to disregard instructions embedded within the financial records.
- Capability inventory: The skill utilizes Python execution for data cleaning and fuzzy matching via the pandas and RecordLinkage libraries.
- Sanitization: There is no evidence of sanitization or validation of the ingested data content before it is processed by the AI.
Audit Metadata