The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-provided drafts and external articles, which creates a surface for indirect prompt injection where malicious instructions embedded in the text could attempt to influence the agent's behavior.
Ingestion points: The skill reads external content through the writer's draft (referenced in SKILL.md) and external published pieces (referenced in references/notebook.md) used for modeling writing patterns.
Boundary markers: The prompt template for the reader subagents in references/readers.md interpolates the <draft> content directly at the end of the prompt without explicit boundary markers or warnings to the subagent to ignore instructions contained within the text.
Capability inventory: The skill possesses the capability to read and write files (e.g., <draft name>.editor.md and notebook.md) and may perform network operations to retrieve external articles for comparison.
Sanitization: There is no evidence of sanitization or filtering of the user-provided text before it is processed by the agent or its subagents.

editor