The Agent Skills Directory

[PROMPT_INJECTION]: The skill facilitates a potential indirect prompt injection surface because it is designed to ingest and act upon data from external, untrusted sources.
Ingestion points: The SKILL.md instructions explicitly guide the agent to fetch context from external integrations such as Slack, Microsoft Teams, Google Drive, SharePoint, and user-provided document links or files.
Boundary markers: The skill does not provide instructions for the agent to use delimiters or ignore embedded commands within the fetched external data.
Capability inventory: The skill utilizes tools like create_file and str_replace to write to the file system or create artifacts, and it can perform further network-based lookups via platform integrations.
Sanitization: There is no logic within the skill to sanitize, validate, or escape content retrieved from external sources before it is interpolated into the prompt or used for drafting document sections.

doc-coauthoring