The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute git commands for branch management. It mitigates command injection risks by implementing a strict slug generation policy in references/branch-policy.md that strips all non-alphanumeric characters from Jira issue summaries before they are interpolated into shell commands.
[PROMPT_INJECTION]: The skill processes untrusted data from Jira issue summaries and descriptions, creating a surface for indirect prompt injection.
Ingestion points: Data enters the agent context via the getJiraIssue and getJiraProjectIssueTypesMetadata tools in SKILL.md.
Boundary markers: The skill does not use specific delimiters but utilizes a 'Plan vs Execute' pattern that requires user oversight before state-changing actions are taken.
Capability inventory: The skill uses Atlassian MCP tools for Jira operations and Bash for git operations.
Sanitization: Input used for branch names is sanitized via regex (/[^a-z0-9\s-]/g) as described in references/branch-policy.md.
[SAFE]: The skill relies on official Atlassian MCP tools for Jira interactions, ensuring that authentication and data access are handled through standard, authorized integration channels.

jira