screwfast
Warn
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to clone a repository from an external GitHub account (Eng0AI/screwfast-template) which is not a verified source. Evidence:
git clone --depth 1 https://github.com/Eng0AI/screwfast-template.git .. - [COMMAND_EXECUTION]: The skill uses shell commands to manage files, directories, and project state during the setup and deployment process. Evidence includes:
mv _temp_template/* _temp_template/.* .,rm -rf _temp_template,rm -rf .git,git init, and variousvercelandnetlifydeployment commands. - [DYNAMIC_EXECUTION]: The build step executes a local JavaScript file (
process-html.mjs) that is part of the downloaded repository, representing execution of potentially untrusted downloaded code. Evidence:npm run buildwhich executesnode process-html.mjs. - [INDIRECT_PROMPT_INJECTION]: The skill processes content from an external repository which could contain instructions for the agent. 1. Ingestion points: SKILL.md (via git clone from Eng0AI/screwfast-template). 2. Boundary markers: Absent. 3. Capability inventory: Shell command execution (git, npm, vercel, netlify, rm, mv). 4. Sanitization: Absent.
Audit Metadata