formal-methods-reconciler
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill operates as a high-level reasoning framework for software verification. It does not contain any executable malicious code or unauthorized commands. The instructions focus on structured analysis and tool interaction.
- [INDIRECT_PROMPT_INJECTION]: The skill manages the risk of processing untrusted data (specs, code, logs) by mandating a structured workflow that separates implementation observations from expected claims and uses formal solvers as the source of truth, minimizing the likelihood of malicious data influencing agent behavior.
- [DATA_EXFILTRATION]: There is no evidence of network exfiltration or sensitive file access. All external URLs referenced in the documentation point to trusted academic research platforms (arXiv) or official tool documentation.
- [REMOTE_CODE_EXECUTION]: The skill describes using verification tools but does not instruct the agent to download or run arbitrary code from untrusted remote sources. It relies on the environment's local formal methods toolchain for verification tasks.
Audit Metadata