The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to capture fleeting ideas (which can include links and external snippets) and later organize them into structured thematic notes. This creates an ingestion surface where untrusted data (e.g., content from a link or a copied snippet containing hidden instructions) could influence the agent's behavior during the 'Phase B: Organize and Merge' workflow.
Ingestion points: User input containing fragments, quotes, or links processed in Step 1 (Receive Input).
Boundary markers: None explicitly mentioned in the document templates to prevent the agent from following instructions embedded within the captures.
Capability inventory: The skill uses yuque_create_doc and yuque_update_doc which allows the agent to write processed (and potentially poisoned) content back to the user's Yuque knowledge base.
Sanitization: No explicit sanitization or filtering of the captured content is described before it is summarized and merged into thematic notes.

daily-capture