nmt-upgrade
Fail
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: CRITICALREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPERSISTENCEINDIRECT_PROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill instructs the agent to download a shell script from 'https://nextmovetheory.com/install.sh' and pipe it directly to 'bash' (or 'iex' on Windows). This pattern permits a remote server to execute arbitrary commands on the user's system without any integrity verification or review.
- [EXTERNAL_DOWNLOADS]: The skill fetches scripts and clones a repository from 'https://nextmovetheory.com' and GitHub ('zamesin/Next-Move-Theory-Canon-and-Skills'), which are third-party sources not recognized as trusted vendors.
- [PERSISTENCE]: The skill overwrites blocks in 'CLAUDE.md' and 'AGENTS.md'. These files store persistent instructions for the AI agent; modifying them with remote content allows a remote actor to permanently alter the agent's behavior.
- [INDIRECT_PROMPT_INJECTION]: The skill creates a vulnerability surface by ingesting data from an external repository and injecting it into agent instruction files without sanitization or validation.
- [COMMAND_EXECUTION]: The skill performs shell operations using 'curl', 'bash', 'git', and 'PowerShell' to modify the local file system and execute remote logic.
Recommendations
- HIGH: Downloads and executes remote code from: https://nextmovetheory.com/install.sh - DO NOT USE without thorough review
- AI detected serious security threats
Audit Metadata