latchbio-integration
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill requires the installation of the 'latch' Python package from PyPI to interact with the LatchBio platform infrastructure.
- [COMMAND_EXECUTION]: Provides instructions for using the Latch CLI, including commands for authentication ('latch login'), project initialization ('latch init'), and workflow deployment ('latch register').
- [INDIRECT_PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection by processing external bioinformatics data files and registry records which could contain malicious instructions.
- Ingestion points: Workflow input parameters defined in
@workflowdecorators, specifically arguments of type 'LatchFile', 'LatchDir', and 'sample_id' (references/workflow-creation.md). - Boundary markers: No explicit delimiters or instruction-ignoring warnings are implemented in the provided code patterns for handling untrusted data.
- Capability inventory: The skill possesses capabilities for cloud file system operations, database record updates (latch.registry), and CLI-based deployment (SKILL.md).
- Sanitization: No sanitization or validation logic for external string inputs is demonstrated in the reference examples.
Audit Metadata