The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install the ctx7 package globally via npm install -g ctx7@latest and facilitates the installation of coding skills from arbitrary GitHub repositories using ctx7 skills install /owner/repo.
[COMMAND_EXECUTION]: The skill's functionality is built around executing shell commands (ctx7). It performs system-level tasks such as writing configuration files to agent directories (e.g., ~/.claude/skills, .mcp.json) and configuring MCP servers during the ctx7 setup process.
[DATA_EXFILTRATION]: The ctx7 skills suggest command reads local project files including package.json, requirements.txt, pyproject.toml, Cargo.toml, go.mod, and Gemfile to identify dependencies and suggest relevant skills from the registry.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it ingests untrusted external data.
Ingestion points: Data enters the agent context through ctx7 docs and ctx7 library (references/docs.md), and ctx7 skills install (references/skills.md).
Boundary markers: Absent. The documentation does not specify the use of delimiters or instructions to ignore embedded commands in the fetched content.
Capability inventory: The CLI has the capability to execute shell commands, write files to the filesystem, and modify agent configurations (references/setup.md).
Sanitization: Absent. While the documentation advises users against including sensitive data in queries, there is no evidence of automated sanitization or validation of the content returned from external sources.

context7-cli