drupal-contrib-mgmt
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted content from the public Drupal.org issue queue and RSS feeds.
- Ingestion points: Processes issue descriptions, comments, and patch files from drupal.org as part of the triage and research workflow (e.g., references/issue-queue-rss-feeds.md).
- Boundary markers: Lacks explicit delimiters or instructions to ignore potentially malicious directives within ingested community content.
- Capability inventory: Possesses capabilities including writing to configuration files (composer.json), executing shell commands via composer and drush, and applying code patches to the filesystem.
- Sanitization: No explicit sanitization of remote data is implemented before the agent processes or acts upon the information.
- [EXTERNAL_DOWNLOADS]: Fetches patch files and project metadata from official Drupal community infrastructure (drupal.org and drupalcode.org) to support module maintenance.
- [COMMAND_EXECUTION]: Executes standard development and maintenance commands such as composer, drush, and git to manage the Drupal project environment.
- [REMOTE_CODE_EXECUTION]: Downloads and applies code patches from external URLs in the Drupal ecosystem to implement bug fixes and compatibility updates.
Audit Metadata