The Agent Skills Directory

[COMMAND_EXECUTION]: The skill interacts with the platform's environment by executing MCP tools (suites_list, tests_search, etc.) and invoking the sync-cases skill to manage testing artifacts.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted data from multiple external sources including Jira, Confluence, Figma, and project source code.
Ingestion points: External context is gathered in SKILL.md (Step 1) from task trackers, requirements documents, and design mockups.
Boundary markers: The skill implements a strong iterative workflow (Steps 1.1, 2, 3, and 4.1) requiring explicit user approval before proceeding between analysis, role selection, and generation phases.
Capability inventory: The skill has the capability to read local source code and write new *.test.md files to the filesystem.
Sanitization: The skill relies primarily on the multi-step human-in-the-loop approval process to sanitize or filter potentially malicious instructions embedded in the source data.

generate-cases