The Agent Skills Directory

[SAFE]: The skill involves the ingestion and processing of untrusted data from external sources, including emails, LinkedIn profiles, and web search results. While this provides a surface for indirect prompt injection, it is essential for the skill's primary function as a negotiation coach.
Ingestion points: External data enters through WebSearch, WebFetch, and various MCP connectors (e.g., Gmail, Slack, CRM) as described in references/context-intake.md.
Boundary markers: The skill does not implement specific boundary markers or delimiters for untrusted content.
Capability inventory: The agent is authorized to use Write, Edit, and Bash(git:*) tools.
Sanitization: The instructions do not mandate sanitization or escaping of external content before processing.
[SAFE]: The Bash tool is explicitly restricted to git commands, mitigating the risk of arbitrary command execution or system-level exploitation.
[SAFE]: The skill's networking operations via WebSearch and WebFetch are limited to the stated purpose of stakeholder research and competitive intelligence. No evidence of data exfiltration to unauthorized third-party domains was found.

influence-and-negotiation