The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructions present a vulnerability surface for indirect prompt injection by combining external data ingestion with state-modifying capabilities.
Ingestion points: The rag_read_full_page and rag_search_knowledge_base tools defined in SKILL.md allow the agent to fetch and process content from external, potentially untrusted sources.
Boundary markers: The instructions lack explicit boundary markers or directives for the agent to treat external content as data rather than instructions.
Capability inventory: The skill provides several write-capable tools in SKILL.md, including manage_task, manage_project, and manage_document, which allow the agent to create, update, and delete records.
Sanitization: There is no requirement or instruction for the agent to sanitize or validate data retrieved via RAG before using it to influence its actions within the task management system.

archon-workflow