conda-recipe
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions for running local build and linting commands including
bioconda-utils build,bioconda-utils lint, andconda mambabuild. These are standard tools for generating and testing software packages in the bioconda ecosystem. - [EXTERNAL_DOWNLOADS]: The skill references downloading source code from well-known services like PyPI and GitHub during the recipe build process. These downloads are standard operations for building packages from source.
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the creation and processing of conda recipe files (
meta.yaml) which include build instructions that are executed by the environment. - Ingestion points: The agent processes recipe files located at
recipes/<package-name>/meta.yamland associatedbuild.shscripts. - Boundary markers: None specified; the skill operates under the assumption that the user provides legitimate recipe definitions for local building.
- Capability inventory: The skill utilizes file system access to write recipes and executes package installation commands (
pip install) and build environments (bioconda-utils,conda-build). - Sanitization: The skill recommends using
bioconda-utils lintto validate the structure and content of recipes before attempting to build them.
Audit Metadata