The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install the ctx7 CLI via NPM (npm install -g ctx7@latest) and enables the downloading of additional skills from GitHub repositories (e.g., /owner/repo). These are core functionalities for the tool's purpose as a documentation and skill manager.
[COMMAND_EXECUTION]: The skill's operation is primarily based on executing ctx7 CLI commands to fetch data, search registries, and perform configuration tasks such as ctx7 setup and ctx7 login.
[DATA_EXFILTRATION]: While the skill communicates with the Context7 API, it includes specific warnings for the agent to avoid including sensitive data like API keys, credentials, or proprietary code in documentation search queries.
[PROMPT_INJECTION]: The skill enables the ingestion of external documentation and third-party skills. This presents a surface for indirect prompt injection, a known characteristic of tools that ingest external content to influence agent behavior. This surface is mitigated by explicit usage warnings.
[REMOTE_CODE_EXECUTION]: The ctx7 skills install command downloads markdown files containing instructions from remote repositories and installs them into agent-monitored directories. While these are instructions rather than binary code, they are designed to be followed by the agent, representing the intended purpose of the system.

context7-cli