antinet-provenance
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or security risks were identified in the skill's instructions or scripts.
- [DATA_EXPOSURE]: The skill records execution traces into local JSONL and JSON files. These operations are limited to the skill's own data directories, and no access to sensitive system configuration or user credentials was found.
- [INDIRECT_PROMPT_INJECTION]: The skill provides an interface for logging events. It includes schema validation to discard malformed events, reducing the surface for injection via audit data. This data is intended for auditing and visualization rather than direct re-execution by the agent.
- [COMMAND_EXECUTION]: The execution script coordinates internal agent stages through the AgentSession class. It does not invoke arbitrary system commands or shell environments with untrusted parameters.
Audit Metadata