cursor-check-agent-compatibility
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from an external directory to perform compatibility and structural scans.
- Ingestion points: The scripts
cursor_plugin_scanner_adapters.pyandcursor_plugin_submission_audit.pyread and parse.cursor-plugin/plugin.json,README.md, and other Markdown files within a user-specified directory. - Boundary markers: The
_boundedfunction inscripts/cursor_plugin_submission_audit.pyimplements path resolution to prevent directory traversal attacks by ensuring all file access remains within the provided plugin root. - Capability inventory: The skill is limited to static analysis (regex, JSON parsing, and globbing); it lacks the capability to execute plugin components, spawn subprocesses from the data, or perform network exfiltration.
- Sanitization: Employs JSON schema validation and strictly bounded file path logic to mitigate malicious data impacts.
- [DATA_EXPOSURE_AND_EXFILTRATION]: The script
scripts/cursor_plugin_scanner_adapters.pycontains a list of credential-shaped keys (e.g.,CURSOR_API_KEY,GITHUB_TOKEN). Review of the logic confirms these are used defensively to identify and block sensitive information from appearing in reports, rather than for harvesting purposes. No network components are present to facilitate exfiltration.
Audit Metadata