The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied feature descriptions, titles, and Figma links to generate structured Jira epics and child tasks.
Ingestion points: The skill is triggered by user requests to create epics or large features, ingesting arbitrary user text into the Epic Template and Suggested Tasks sections.
Boundary markers: The skill uses simple curly brace placeholders (e.g., {Epic Title}, {overview}) but lacks explicit boundary markers or instructions to the model to ignore potential injection attempts within the user-provided content.
Capability inventory: The skill utilizes mcp-atlassian__jira_create_issue, mcp-atlassian__jira_update_issue, and mcp-atlassian__jira_transition_issue to write data to an external Jira instance.
Sanitization: No sanitization or validation of the user input is described; the instructions focus on formatting transitions between Markdown and Jira Wiki markup.

jira-epic