fw-review
Pass
Audited by Gen Agent Trust Hub on May 6, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/fdk-errors-warnings.jsexecutes thefdk validatecommand to leverage official platform validation logic. This behavior is expected for an application audit tool. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted code and configuration data from the local workspace. Ingestion points include
manifest.json,config/iparams.json, and source code files. Boundary markers are not explicitly defined in the instructions. Capability inventory includes file system access and execution of thefdktool. Sanitization is performed through deterministic regex analysis in the included scripts. This represents a low-risk surface inherent to the skill's primary function. - [SAFE]: The skill proactively checks for security issues, including hardcoded OAuth credentials (
scripts/oauth-config-usage.js) and non-allowlisted external import sources (scripts/external-import-sources.js).
Audit Metadata