The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill facilitates the installation of third-party agent skills by cloning GitHub repositories and copying configuration files (SKILL.md) into local application directories (e.g., ~/.claude/skills/). This pattern allows remote content to define future agent behavior on the host system.- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it automatically processes and installs content from external GitHub repositories.
Ingestion points: Metadata and skill content are fetched from remote repositories via the github-skills API.
Boundary markers: No protective delimiters or warnings are used when the agent handles the content of the external skills during installation.
Capability inventory: The skill uses curl, cp, and shell execution to manipulate the file system and network.
Sanitization: The skill does not implement any validation or sanitization of the downloaded instructions before installation.- [COMMAND_EXECUTION]: Performs several shell-based operations, including global package management (npm install -g omniroute), local API interaction via curl, and directory inspection using ls.- [EXTERNAL_DOWNLOADS]: Downloads the omniroute CLI tool from npm and fetches skill data from GitHub. Both npm and GitHub are well-known services.

cli-skill-collector