tracking-service-reliability
Warn
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: MEDIUMDYNAMIC_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DYNAMIC_EXECUTION]: The script
scripts/generate_sla_report.pyimplements agenerate_scriptfunction that dynamically constructs shell script files (.sh) from templates and content passed via command-line arguments. - [COMMAND_EXECUTION]: Dynamically created shell scripts are automatically granted executable permissions using
chmod(0o755)inscripts/generate_sla_report.py, allowing the system to run the generated code directly. - [INDIRECT_PROMPT_INJECTION]: The skill acts as a processing pipeline for external service reliability data which poses an injection risk.
- Ingestion points: External data is ingested via the
--contentand--configparameters of the report generation script inscripts/generate_sla_report.py. - Boundary markers: The implementation lacks delimiters or protective instructions to distinguish between data and potentially malicious embedded instructions.
- Capability inventory: The skill possesses file-write and permission-modification capabilities through its Python utility script.
- Sanitization: There is no validation or escaping logic applied to inputs before they are incorporated into executable scripts.
Audit Metadata