The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from user conversations and external support cases to populate Jira issue fields.
Ingestion points: Processes user input in workflows/create-issue.md and support case data in references/support-intake.md.
Boundary markers: No explicit delimiters are used for untrusted content, but the workflow includes a mandatory user review step in Step 6.
Capability inventory: The skill can create, update, and link issues in Jira via the acli tool.
Sanitization: Explicitly mandates the removal of customer-identifying information (names) from public fields in projects like RHDHBUGS and RHDHPLAN.
[COMMAND_EXECUTION]: The skill uses the Atlassian CLI (acli) and standard shell utilities for its operations.
Executes acli commands to interact with Jira APIs for issue creation, linking, and commenting.
Utilizes mktemp and cat to generate temporary files for user review, ensuring a write-gate is respected before any external state is modified.

rhdh-jira-create