The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates legal workflows by pre-filling and invoking other internal agent commands like /matter-intake, /demand-intake, and /legal-hold. This behavior is consistent with its role as a workflow orchestrator for legal matters.
[EXTERNAL_DOWNLOADS]: The skill may access external information through web searches if primary legal research tools are insufficient. This capability is explicitly gated by a user confirmation step, and the instructions require the agent to warn the user that such information requires manual verification.
[PROMPT_INJECTION]: The skill processes untrusted demand letters provided by users, which represents an indirect prompt injection surface.
Ingestion points: The skill reads external file contents or session documents as the primary step in its workflow.
Boundary markers: No explicit delimiters are specified in the instructions to isolate the document content from the agent's logic.
Capability inventory: The skill has permissions to write analysis files (triage.md), copy/link files in local directories, and trigger further agent commands.
Sanitization: The skill relies on a structured extraction template and categorical evaluation rather than active content sanitization, but it limits the agent's role to triage and routing rather than automated execution of document-derived instructions.

demand-received