browser-native
Pass
Audited by Gen Agent Trust Hub on Aug 10, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill's primary function is to read and analyze local package.json files. All logic is contained within local scripts, and no external dependencies or network requests are utilized.
- [COMMAND_EXECUTION]: The skill uses a local Node.js script to perform its analysis. The CLI tool supports reading from and writing to the local filesystem (for report generation), which are standard and expected operations for this type of utility.
- [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection through the ingestion of external package.json files during the scanning process.
- Ingestion points: Manifest files are discovered and read in scripts/scanner.js using the node:fs/readFileSync method.
- Boundary markers: None. The tool processes JSON keys directly without specific warnings to the agent regarding potentially malicious content embedded in the manifest data.
- Capability inventory: The skill can write report output to a user-specified file path using node:fs/writeFileSync in scripts/cli.js.
- Sanitization: Input keys extracted from the package manifest are checked against a static, hardcoded database in scripts/replacements.js, which restricts the potential for malicious data to influence the tool's behavior or output beyond the content of the report itself.
Audit Metadata