bitwarden-browser-aac-login

Pass

Audited by Gen Agent Trust Hub on Sep 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions involve executing the aac (Bitwarden Agent Access) CLI to manage vault connections and run processes with sensitive credentials (AAC_USERNAME, AAC_PASSWORD) injected into the child environment.
  • [DYNAMIC_EXECUTION]: The recommended workflow generates and executes a temporary Node.js helper script via node -e at runtime to transfer credentials from the environment to a local loopback server for browser automation.
  • [INDIRECT_PROMPT_INJECTION]: The skill interacts with untrusted external website content to identify login forms and verify login success, creating a surface for potential instruction injection. • Ingestion points: Browser-control tools read DOM elements, page titles, and error messages from target websites as described in SKILL.md and references/browser-bridge.md. • Boundary markers: The instructions focus on finding specific form locators but do not provide explicit delimiters to prevent the agent from following instructions embedded in the page content. • Capability inventory: The skill has access to shell command execution (aac), browser automation, and local socket creation for internal communication. • Sanitization: The skill contains detailed instructions in the 'Sanitization' section to redact credentials, tokens, and identifiers from all output and reports.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 10, 2026, 08:55 AM