The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external sources provided by the user, which could contain instructions intended to deviate the agent from its intended workflow.
Ingestion points: Fetches shared documents via links or reads user-provided template files during the 'Context Gathering' stage in SKILL.md.
Boundary markers: The instructions do not specify the use of delimiters (e.g., XML tags or triple quotes) or explicit warnings for the agent to ignore instructions embedded within the processed document content.
Capability inventory: The skill uses create_file and str_replace to modify the workspace, can perform network operations via integrations (Slack, Google Drive, etc.), and can invoke sub-agents for 'Reader Testing' as defined in SKILL.md.
Sanitization: There is no mention of sanitizing, escaping, or validating the content of external documents before it is used to influence the drafting process.

doc-coauthoring