context7-auto-research
Warn
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
- [COMMAND_EXECUTION] (MEDIUM): Risk of shell command injection via string interpolation. The instructions guide the agent to place user-controlled queries directly into bash commands such as 'node context7-api.cjs search "" ""'. This allows an attacker to execute arbitrary commands if the user input contains shell metacharacters.
- [PROMPT_INJECTION] (LOW): Indirect prompt injection surface detected. The skill fetches documentation from context7.com and integrates it into the agent's context without boundary markers or sanitization. 1. Ingestion point: Output from context7-api.cjs. 2. Boundary markers: Absent. 3. Capability inventory: Bash, Task, Read, Write. 4. Sanitization: Absent.
- [REMOTE_CODE_EXECUTION] (MEDIUM): The skill relies on an external Node.js script 'context7-api.cjs' which is executed via the Task tool but is not provided in the skill package for security auditing.
- [EXTERNAL_DOWNLOADS] (LOW): The skill communicates with an external API (context7.com) which is not on the trusted whitelist. This is the primary purpose of the skill, but it introduces an external dependency on untrusted data.
Audit Metadata