worktree-manager

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • Indirect Prompt Injection (SAFE): The skill ingests Git metadata (branch names and paths) from command output, which is a theoretical surface for indirect prompt injection. However, this is inherent to the skill's primary purpose and no specific vulnerabilities were detected.\n
  • Ingestion points: Output from git worktree list as described in references/listing.md.\n
  • Boundary markers: Absent; command output is processed without specific delimiters or instructions to ignore embedded content.\n
  • Capability inventory: Git worktree management (add, remove, prune), directory deletion (rmdir), and .gitignore updates via echo.\n
  • Sanitization: Absent; the skill relies on the agent's default handling of command arguments.\n- Command Execution (SAFE): All commands (git, grep, echo, rmdir, basename) are localized to the repository and used according to best practices for Git worktree management.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:29 PM