The Agent Skills Directory

[PROMPT_INJECTION]: The skill functions as an ingestion point for untrusted data from Jira issues, presenting a surface for indirect prompt injection where malicious instructions embedded in ticket content could attempt to influence the agent's behavior.
Ingestion points: The skill retrieves external content from Jira via the getJiraIssue (MCP) and issue view (CLI) operations documented in SKILL.md and references/mcp.md.
Boundary markers: The skill instructions in SKILL.md explicitly mandate that the agent must show proposed changes and wait for user approval before execution, acting as a human-in-the-loop boundary.
Capability inventory: The skill possesses the ability to modify Jira tickets, transition workflows, and execute shell commands via the jira CLI.
Sanitization: Mandatory safety protocols require the agent to fetch the current issue state first and show original descriptions to the user to prevent state-manipulation or unauthorized content replacement.
[COMMAND_EXECUTION]: The skill utilizes the jira CLI for its primary functionality. It includes a specific implementation pattern in references/commands.md for handling multi-line strings by writing to a temporary file in /tmp before execution, which is a standard operational procedure for this tool and does not represent an elevated security risk in this context.

jira