bioinformatics-project

Pass

Audited by Gen Agent Trust Hub on Sep 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The scripts/scaffold_project.py script creates directories and writes multiple files to the local file system using the pathlib and os libraries. It also programmatically sets the executable bit on generated scripts (like runall) using chmod.
  • [INDIRECT_PROMPT_INJECTION]: The scaffolding script ingests user-provided strings via command-line arguments and interpolates them into the content of generated files, which could be exploited if untrusted data is passed to the script.
  • Ingestion points: User input provided to --name, --objective, and --author in scripts/scaffold_project.py is written directly into README.md, pixi.toml, LICENSE, and CITATION.cff.
  • Boundary markers: None are used when interpolating these values into the markdown or configuration files.
  • Capability inventory: The skill has the capability to write and modify files at arbitrary paths specified by the user via the project_dir argument.
  • Sanitization: There is minimal sanitization; the project name is normalized to a safe slug for pixi.toml via pixi_name(), and quotes are replaced in CITATION.cff, but the objective string is written to README.md largely as-is.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 1, 2026, 03:18 PM
Security Audit — agent-trust-hub — bioinformatics-project