pci-secure-software
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill follows a 'fail-closed' security philosophy. It uses a deterministic, non-LLM citation verifier (
tools/pci-sss/citation_verify.py) to confirm that any code snippets cited as evidence actually exist in the analyzed codebase, effectively mitigating the risk of LLM hallucinations. - [SAFE]: Implements a blind adversarial refutation process where independent agents attempt to disprove findings, ensuring that compliance verdicts (MET/NOT_MET) are rigorously tested before inclusion in the final report.
- [COMMAND_EXECUTION]: Benign execution of local Python and JavaScript scripts for core logic, including applicability filtering, aggregation, and report generation. These operations are confined to the engagement output directory and are standard for the skill's intended purpose.
- [EXTERNAL_DOWNLOADS]: References official PCI SCC documentation and utilizes trusted sub-skills from the Transilience AI vendor (e.g.,
skills/source-code-scanning,skills/cryptography) for specialized analysis tasks.
Audit Metadata