The Agent Skills Directory

[PROMPT_INJECTION]: The skill facilitates a workflow where the agent processes external research data, establishing an attack surface for indirect prompt injection. * Ingestion points: External content gathered during the research and professor-critic iterative process. * Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions for separating untrusted research data from agent instructions. * Capability inventory: The skill utilizes task management tools (TaskCreate, TaskUpdate) and file system writes to the .omw/ directory. * Sanitization: No specific sanitization or validation of the ingested research content is described before it influences task updates or file writes.

autoresearch-goal