The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes several Python scripts to interface with the Atlassian CLI (acli). Scripts such as setup.py, parse_issues.py, validate_components.py, and validate_field_ids.py use subprocess.run to execute commands. These executions use list-based arguments, which is a secure practice that avoids shell interpretation and mitigates command injection risks. The scripts search for the acli executable in standard system paths and user-local directories.
[INDIRECT_PROMPT_INJECTION]: The skill processes Jira issue data, including summaries and descriptions, which may be authored by third parties. This creates a surface for indirect prompt injection where instructions embedded in Jira content could influence the agent's behavior.
Ingestion points: Scripts such as parse_issues.py and adf_milestones.py ingest JSON data from Jira and process text from fields like 'summary' and 'description'.
Boundary markers: The skill documentation provides clear guidance on the intended read-only scope, but the underlying scripts do not use specific delimiters or markers to isolate untrusted text.
Capability inventory: The skill is capable of performing bulk searches and viewing detailed issue data using acli. It does not perform writes or network operations outside of the authenticated CLI tool.
Sanitization: No explicit sanitization or filtering is performed on the text content extracted from Jira issues before it is presented to the agent.

rhdh-jira-api