contract-performance-bond-check-free
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted financial ledger data provided by the user via the command-line interface. This data is parsed and then presented back to the AI agent as an audit report. This ingestion of external data constitutes an attack surface where instructions embedded in the ledger fields (such as project names or notes) could attempt to influence the AI agent's subsequent behavior.
- Ingestion points: The
scripts/run.mjsscript reads content from external files provided via the--inputargument. - Boundary markers: While the logic returns structured results, the skill does not wrap the processed untrusted input in specific delimiters or provide 'ignore' instructions to the agent's context to mitigate potential injection.
- Capability inventory: The skill environment is configured with
ReadandBashtools, which are used to execute the local analysis scripts. - Sanitization: The execution engine in
scripts/engine/contract-performance-bond-check.jsperforms strict numeric and date normalization (e.g., usingnormNumberandnormDate), which helps validate individual data points and reduces the risk of non-textual injection patterns.
Audit Metadata