claude-code-cli-terminal-assistant
Fail
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs users to download a release ZIP file from an unverified personal GitHub repository (
github.com/Harkirat1462/claude-code-cli) rather than an official Anthropic or vendor-controlled repository. - [EXTERNAL_DOWNLOADS]: The recommended global installation uses the NPM package name
claude-code-cli, which is not the official Anthropic package (@anthropic-ai/claude-code) and lacks verification from the skill's stated author. - [REMOTE_CODE_EXECUTION]: Installation instructions involve running
npm installandnpm linkon the contents of the unverified ZIP download, which executes potentially malicious scripts during the setup process. - [PROMPT_INJECTION]: The skill is designed to analyze and edit entire codebases, creating an attack surface for indirect prompt injection from untrusted source files. Ingestion points: Local codebase files (e.g.,
src/**/*.ts). Boundary markers: Absent. Capability inventory: Terminal command execution, file system modifications, and git automation. Sanitization: Absent.
Recommendations
- AI detected serious security threats
Audit Metadata