vhs
Pass
Audited by Gen Agent Trust Hub on Mar 6, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill guides the agent in handling
.tapefiles, which contain terminal command sequences. This presents a surface for indirect prompt injection if an agent processes files from untrusted sources without sanitization.\n - Ingestion points: Tasks involving
.tapefiles as defined inSKILL.md.\n - Boundary markers: None identified.\n
- Capability inventory: The skill describes terminal command execution (
Type,Enter), keystroke simulation (Ctrl,Key Repeat), and environment variable setup (Env).\n - Sanitization: No specific content sanitization or isolation protocols are suggested.\n- [COMMAND_EXECUTION]: Documentation provides instructions for executing the
vhsCLI and installing dependencies via system package managers.\n- [EXTERNAL_DOWNLOADS]: The skill references established GitHub Actions (charmbracelet/vhs-action,stefanzweifel/git-auto-commit-action) which are well-known and widely used services.
Audit Metadata