The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill clones a software repository from an external source (github.com/nlweb-ai/NLWeb) and fetches multiple documentation files using WebFetch.
[REMOTE_CODE_EXECUTION]: The instructions direct the agent to execute setup tools and initialization scripts bundled within the cloned repository, such as nlweb init-python and nlweb init.
[COMMAND_EXECUTION]: The skill uses the Bash tool to perform environment configuration, including creating virtual environments and installing Python packages from a requirements file (pip install -r requirements.txt).
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it requires the agent to fetch and process live documentation and release notes from a remote repository to determine configuration parameters and environment variables.
Ingestion points: SKILL.md (WebFetch of README.md, release notes, and CLI documentation from github.com/nlweb-ai/NLWeb).
Boundary markers: Absent; no delimiters are used to separate remote content from instructions.
Capability inventory: SKILL.md (Bash tool for cloning, package installation, and execution of the nlweb CLI; Write tool for creating .env and YAML config files).
Sanitization: Absent; the agent is not instructed to validate or sanitize content retrieved from the remote repository before acting on it.

nlweb-setup