bridge-plugin

Pass

Audited by Gen Agent Trust Hub on Mar 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The install_all_plugins.py script automates the deployment of multiple plugins by executing the bridge_installer.py script via subprocess.run. This is a standard architectural pattern for batch operations within the local environment and does not involve executing arbitrary shell commands or untrusted input.
  • [EXTERNAL_DOWNLOADS]: The skill specifies a dependency on the yaml Python package via pip in its metadata. No other external network requests, remote script executions, or unauthorized downloads were found.
  • [PROMPT_INJECTION]: The skill serves as a bridge for external plugin data, which creates a surface for indirect prompt injection if the source plugins are compromised.
  • Ingestion points: Plugin content is read from the plugins/ directory and processed in bridge_installer.py.
  • Boundary markers: The transformation logic does not implement boundary markers or instructions to ignore embedded commands within the bridged content.
  • Capability inventory: The skill possesses capabilities for file system modification (shutil.copytree, Path.write_text) and local script execution via subprocess.run.
  • Sanitization: The script performs functional transformations like actor flag swapping but does not validate or sanitize the instructional content of the plugins being processed.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 14, 2026, 08:58 AM