The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The documentation describes an installation process using npx skills add StanfordSpezi/SpeziVibe. This fetches code from the StanfordSpezi organization's repository. As this originates from the skill's official vendor infrastructure, it is a standard deployment mechanism.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and synthesize untrusted external data (such as patient interviews, research papers, and clinical observations) into a persistent wiki, creating a vulnerability surface where malicious instructions in those sources could influence agent behavior.
Ingestion points: External content is introduced through file uploads, text pastes, or URLs into the wiki/raw/ directory (referenced in SKILL.md).
Boundary markers: The instructions lack specific technical delimiters or structured formatting requirements to isolate ingested text from the agent's processing logic.
Capability inventory: The agent has permissions to write and modify multiple files within the repository, including wiki/pages/ and the core configuration files AGENTS.md or CLAUDE.md (referenced in SKILL.md).
Sanitization: The skill relies on natural language instructions to 'flag contradictions' rather than technical sanitization or filtering of the external input.

project-wiki