skills/hussainweb/skills/beads-init/Gen Agent Trust Hub

beads-init

Pass

Audited by Gen Agent Trust Hub on Nov 3, 2026

Risk Level: SAFECOMMAND_EXECUTIONPERSISTENCEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill makes extensive use of the Bash tool to execute bd and git commands for repository initialization, configuration management, and environment setup.
  • [PERSISTENCE]: The skill installs git hooks into the .git/hooks/ directory and configures SessionStart hooks for AI agents (e.g., in .claude/settings.json). These mechanisms are used to automate issue tracker synchronization and tool priming but represent a persistent code execution surface.
  • [EXTERNAL_DOWNLOADS]: The bd bootstrap functionality involves cloning a Dolt database from the repository's git origin. This results in the download of external data from a remote source defined by the project's configuration.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests external data from the repository environment, such as directory names and git remote information, to dynamically determine configuration values like issue prefixes. This data is interpolated into shell commands. The risk is mitigated by instructions requiring the agent to propose values to the user and request confirmation for non-standard inputs.
  • [DATA_EXPOSURE]: The skill reads local git configurations and issue tracker settings to audit the state of the repository. No evidence of data exfiltration to unauthorized external domains was detected.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 3, 2026, 01:05 AM