sf-hxl
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The
tools/check_sources.pyscript performs network requests todeveloper.salesforce.comto check for documentation drift. This targets a well-known, trusted service and is used for maintenance purposes. - [COMMAND_EXECUTION]: The test suite in
tests/test_hxl.pyutilizessubprocess.runto execute the bundledscripts/scaffold.pyscript during regression testing. This is a standard testing practice using the local Python interpreter and bundled components. - [INDIRECT_PROMPT_INJECTION]: The skill provides tools for processing Salesforce project metadata, which constitutes a theoretical attack surface.
- Ingestion points:
scripts/validate.pyandscripts/scaffold.pyread JSON and XML files from user-specified project directories (e.g.,sfdx-project.json,schema.json). - Boundary markers: The tools enforce strict metadata schemas for HXL widgets and Lightning types, rejecting malformed or unknown structures.
- Capability inventory:
scripts/scaffold.py(writes metadata files),tools/install_skill.py(copies files to local skills directory). - Sanitization: API identifiers and field names are validated against regex patterns (
[A-Za-z][A-Za-z0-9_]*) to prevent path traversal and command injection.
Audit Metadata