academic-paper-verify
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, such as credential harvesting, data exfiltration, or obfuscation, were identified. The skill's behavior is entirely consistent with its stated purpose of verifying research replicability.
- [COMMAND_EXECUTION]: The skill directs the agent to scan local project directories and execute R scripts to inventory project assets and audit analytical pipelines. These operations are restricted to the user's local environment.
- [REMOTE_CODE_EXECUTION]: The skill generates and runs a local verification script (
tests/verify_replication.R) based on an included R template. This automated execution is the core feature of the replication check phase and is performed on the user's own project files, posing no risk of unauthorized external access.
Audit Metadata