make-a-woke-profile-with-chatgpt

Pass

Audited by Gen Agent Trust Hub on Nov 14, 2026

Risk Level: SAFECREDENTIALS_UNSAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [CREDENTIALS_UNSAFE]: The script scripts/wokify.py attempts to read an API key from ~/.varg/credentials, which is a standard location for vendor-specific credentials for the varg.ai service.
  • [COMMAND_EXECUTION]: The scripts scripts/collage.py and scripts/presenter.py use subprocess.run to execute ffmpeg commands for generating video and GIF content. The parameters are mostly static but rely on file paths generated or provided during execution.
  • [REMOTE_CODE_EXECUTION]: The scripts/presenter.py script generates a package.json file and provides instructions to run npx hyperframes@0.9.15 for rendering video content. This involves downloading and executing code from a remote registry.
  • [EXTERNAL_DOWNLOADS]: The skill configuration and scripts involve downloading assets from s3.varg.ai and executing remote packages via npx.
  • [INDIRECT_PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it takes user-supplied name and pronouns via CLI arguments in scripts/wokify.py and interpolates them directly into prompts for the gpt_image_2_6 model without sanitization.
  • Ingestion points: --name and --pronouns CLI arguments in wokify.py.
  • Boundary markers: None present in the prompt templates.
  • Capability inventory: The skill has the ability to write files, execute shell commands (via ffmpeg), and perform network requests to the varg.ai API.
  • Sanitization: No escaping or validation is performed on the user-supplied strings before prompt construction.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 14, 2026, 09:50 PM