The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted external data (meeting notes, requirements, code context) to generate Jira tickets, creating a surface for indirect prompt injection.
Ingestion points: External content enters the agent's context through user-provided source material, such as requirements, code context, and meeting notes as described in SKILL.md.
Boundary markers: The instructions implement a safety check requiring the agent to 'preview the payload' and 'obtain confirmation' before executing the ticket creation in SKILL.md.
Capability inventory: The skill has the capability to write to external Jira instances using the mcp__atlassian__createJiraIssue tool.
Sanitization: There are no explicit instructions for automated sanitization or data escaping, relying primarily on human-in-the-loop review to mitigate malicious input.

jira-mcp