How this skill is triggered — by the user, by Claude, or both
Slash command
/workflow-commands:continueThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Resume work on an existing task in a new session. Recovers context from beads and sets up for continued work.
Resume work on an existing task in a new session. Recovers context from beads and sets up for continued work.
Use when: Starting a new session and continuing previous work.
Announce: "I'm resuming work on this task."
<task-id>
beads-a1b2 or just a1b2)If no task ID provided, show available work:
bd ready
bd list --status=in_progress
# Get task details
bd show <task-id>
Display to user:
If not already in_progress:
bd update <task-id> --status in_progress
Log continuation:
deciduous add action "Continuing: <task title>"
Present the user with:
## Resuming: <task title>
**Status:** <status>
**Priority:** <priority>
### Context
<description and notes from task>
### Recent Activity
<comments, especially LEARNED notes>
### Dependencies
- Blocked by: <list or "none">
- Blocking: <list or "none">
### Suggested Next Steps
<based on task type and state>
Continue working on the task. Use normal patterns:
# Discoveries
bd comment <task-id> "LEARNED: <finding>"
# Decisions
deciduous add decision "<choice> because <rationale>" -c 80 # -c takes integer 0-100, NOT float
# Progress
bd comment <task-id> "PROGRESS: <what was done>"
bd update <task-id> --status done
deciduous add outcome "Completed: <summary>"
bd close <task-id> --reason "Complete"
If you don't remember the task ID:
# Show in-progress work
bd list --status=in_progress
# Show ready work (no blockers)
bd ready
# Search by text
bd search "<keyword>"
Direct resume:
/continue beads-a1b2
Find and resume:
User: /continue
Claude: Here's your current work:
In Progress:
- beads-a1b2: Implement token refresh (High)
Ready to Start:
- beads-c3d4: Add logout endpoint (Normal)
Which task would you like to continue?
npx claudepluginhub kylestratis/kyle-claude-plugins --plugin workflow-commandsCreates bite-sized, testable implementation plans from specs or requirements, with file structure and task decomposition. Activates before coding multi-step tasks.