From Leadtime
Use when the user asks to work with Leadtime, inspect/update Leadtime data, manage tasks, summarize work, or sends a Leadtime link.
How this skill is triggered — by the user, by Claude, or both
Slash command
/leadtime:leadtimeThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Use this skill when the user asks Claude to work with Leadtime, inspect or update tasks/projects/workspace data, plan task work, draft comments, create tasks, summarize reports, or sends a Leadtime link.
Use this skill when the user asks Claude to work with Leadtime, inspect or update tasks/projects/workspace data, plan task work, draft comments, create tasks, summarize reports, or sends a Leadtime link.
Prefer the bundled Leadtime MCP connector for real workspace data. Do not tell the user to create a personal access token for normal in-product work; use the plugin MCP OAuth connection instead. Ask for an API key only when the user wants to build scripts, automations, or third-party integrations outside Claude.
If the user asks for Leadtime data and MCP is not authenticated, tell them to reinstall or upgrade the Leadtime plugin and approve the Leadtime browser OAuth flow. If Claude still does not start the OAuth flow, the fallback is adding the Leadtime MCP server manually with OAuth, but that is a troubleshooting path rather than the default plugin experience.
For MCP/API work, use this sequence:
For vague action requests like "create a task" without details, ask a short clarifying question first.
Many Leadtime fields store ProseMirror editor content. When creating or updating descriptions, comments, journals, product descriptions, document content, or similar rich-text fields:
<p></p> or <p class="paragraph-base"></p> if required.Closed; names like Done or Completed are not enough.Before creating a task, resolve:
Required task fields usually include title, project, type, status, priority, and description. Default priority to Normal and empty description to <p></p> when allowed. If project or title is missing, ask the user.
When the user asks you to create follow-up work for "you", "yourself", "Leadtime agent", or the internal agent, create a normal task assigned to the built-in bot named Leadtime Agent if available. Make that task self-contained: include outcome, constraints, source links, and prior decisions.
For task reports, avoid generic boilerplate. Use actual task details, comments, history, status changes, time, assignee/accountable, blockers, and next steps.
Use this order:
For "my tasks" summaries, resolve the current user, default to open assigned tasks, group by project or status, highlight overdue/high-priority items, and resolve IDs before presenting.
Use full Leadtime URLs when you know the workspace URL. Entity links:
/task/[taskNumber] using singular task; never /tasks/ for a single task./project/[projectId]/overview/organizations/[organizationId]/overview/billing/receivables/new/[invoiceId]/overview/planning/tasksFor one entity, use concise prose or bullets. For two or more rows, use a markdown table with practical columns. Default task columns: task number/link, title, status, type, assignee, project, due date. Avoid raw UUIDs in user-facing output unless explicitly requested.
For implementation work in a Leadtime repository, follow that repository's local instructions and domain docs first.
Guides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.
npx claudepluginhub workcio/claude-leadtime-plugin --plugin leadtime