The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructions define an attack surface for indirect prompt injection.
Ingestion points: In the "Meeting Briefings" section, the agent is instructed to "Gather context from connected sources" including external communications like "Email (recent correspondence)" and "Chat (recent discussions)," as well as "Documents."
Boundary markers: The skill lacks instructions for the agent to use specific delimiters or to disregard potentially conflicting instructions embedded within the external data sources during analysis.
Capability inventory: The skill's workflow involves reading local configuration files (negotiation playbooks), accessing communication logs (email/chat), and synthesizing findings into briefings and templates, providing a pathway for malicious content to influence agent outputs.
Sanitization: No guidelines are provided for validating or sanitizing content retrieved from external sources before it is processed into the legal operations workflow.
[NO_CODE]: The skill consists solely of Markdown-based instructions and response templates for the agent. It does not include any executable scripts, binaries, or automated installation steps, which reduces the risk of direct code-based attacks.

generic-legal-ops