beancount-accounting
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- Prompt Injection (SAFE): No attempts to override safety filters or hijack agent behavior were found.
- Data Exposure & Exfiltration (SAFE): No hardcoded secrets, credentials, or unauthorized network calls were identified. Examples use standard local file paths.
- Obfuscation (SAFE): All content is provided in clear, human-readable markdown without any hidden or encoded segments.
- Unverifiable Dependencies & Remote Code Execution (SAFE): The skill references standard Beancount command-line tools but does not include any remote script downloads or installations from untrusted sources.
- Privilege Escalation (SAFE): No commands related to privilege elevation (like sudo) or system permission modifications were found.
- Persistence Mechanisms (SAFE): No instructions or scripts attempt to establish persistence on the host system.
- Metadata Poisoning (SAFE): The skill metadata accurately describes its purpose as a Beancount accounting assistant.
- Indirect Prompt Injection (SAFE): While the skill involves processing financial data files, the documentation focuses on structured syntax and validation, minimizing the risk of data being misinterpreted as commands.
- Time-Delayed / Conditional Attacks (SAFE): No logic gates or conditional behaviors based on time or environment were detected.
- Dynamic Execution (SAFE): No runtime code generation, unsafe deserialization, or script assembly was identified.
Audit Metadata