The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill depends on several third-party agent skills hosted in GitHub repositories that are not verified as trusted vendors or well-known technology organizations.
Evidence: subagents/preflight-checker-manifest.md lists numerous required skills to be installed via skills install, including resources from obra/superpowers, antfu/skills, softaworks/agent-toolkit, blader/humanizer, sickn33/antigravity-awesome-skills, and wshobson/agents.
[PROMPT_INJECTION]: The skill processes external data from Jira tickets, which represents a surface for indirect prompt injection attacks.
Ingestion points: Data is fetched from Jira and stored in local markdown files (e.g., docs/<KEY>.md) by the fetching-jira-ticket skill in Phase 1 and the ticket-status-checker subagent.
Boundary markers: There are no instructions or patterns defined to wrap ingested ticket content in delimiters or to apply specific 'ignore embedded instructions' markers.
Capability inventory: The orchestrator dispatches execution tasks to downstream skills like executing-jira-task, which has broad capabilities including file system modification, git operations, and code execution.
Sanitization: No sanitization, validation, or filtering of the incoming Jira ticket description, comments, or fields is implemented.

orchestrating-jira-workflow