The Agent Skills Directory

[SAFE]: The skill adheres to security best practices by recommending the use of environment variables for API keys and secrets rather than hardcoding them. No malicious obfuscation, privilege escalation, or persistence patterns were found.
[EXTERNAL_DOWNLOADS]: The skill references several well-known and trusted Python packages, including the official benchling-sdk, biopython, and python-dotenv. These are standard tools for interacting with the Benchling platform and handling biological data.
[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection:
Ingestion points: Data ingested from Benchling notebook entries, entity fields, and workflow tasks (e.g., benchling.entries.get()).
Boundary markers: Not explicitly defined in the provided code templates or prompt instructions.
Capability inventory: Comprehensive CRUD operations on Benchling registry, inventory, and notebook data via the SDK (documented in references/sdk_reference.md).
Sanitization: No explicit sanitization or validation of the retrieved text data is shown in the examples. This is a common architectural pattern for data-driven skills and is typically mitigated by the model's safety layer.

benchling-integration