Stats
Actions
Tags
From agkan-skills
Use when reviewing backlog tasks to assess decomposition, implementation readiness, and priority ordering before development begins.
How this skill is triggered — by the user, by Claude, or both
Slash command
/agkan-skills:agkan-planningThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
A planning workflow that uses agkan to review backlog tasks and make decisions about decomposition, moving to Ready status, and deferring tasks.
A planning workflow that uses agkan to review backlog tasks and make decisions about decomposition, moving to Ready status, and deferring tasks.
CONFIG=$(agkan config get --json 2>/dev/null || echo '{}')
PLANNING_MODEL=$(echo "$CONFIG" | jq -r '.config.models.planning.model // "sonnet"')
PLANNING_EFFORT=$(echo "$CONFIG" | jq -r '.config.models.planning.effort // "medium"')
agkan task list --status backlog --json
For each task, use the Task tool (general-purpose sub-agent) to review.
Do not use Skill("agkan-planning-subtask"). Instead, instruct the sub-agent to read the SKILL.md file directly.
Why SKILL.md path instead of
Skill()? Sub-agents spawned via the Task tool start with a fresh context.Skill()loads skill content into the current conversation, but a sub-agent needs its instructions embedded in its prompt. Providing the SKILL.md path directly in the prompt is the reliable way to pass workflow instructions to a sub-agent.
Task(
subagent_type="general-purpose",
model="<PLANNING_MODEL>",
description="Review task #<id>",
prompt="""
Please review the following backlog task.
## Task Information
- ID: <id>
- Title: <title>
- Body: <body>
## Procedure
Read .claude/skills/agkan-planning-subtask/SKILL.md and follow its procedures to review.
## Important Constraint
Your role is ONLY to review the task and update its status in agkan (e.g., move to ready, decompose, or tag for deferral).
Do NOT implement the task. Do NOT edit any source code or codebase files.
## Effort / Thoroughness
Effort level: <PLANNING_EFFORT>
- low: Quick assessment. Focus on obvious gaps or blockers. Minimal research.
- medium: Standard review. Check requirements clarity, decomposition needs, and implementation readiness.
- high: Thorough review. Deep analysis of dependencies, edge cases, and potential risks. Research codebase as needed.
"""
)
If a task ID is specified by the user, retrieve and review only that target task:
agkan task get <id> --json
Then delegate only that single task to a sub-agent using the same Task call format above.
After the sub-agent completes, re-fetch the backlog task list to pick up any changes (e.g., tasks added by decomposition):
agkan task list --status backlog --json
If there is no instruction to end from the user and backlog tasks remain, select the next task and repeat from step 2.
If no backlog tasks remain, end the session.
Retrieve backlog tasks
↓
Delegate one task to sub-agent
↓
Sub-agent reviews and makes decision
- Content unclear? → Research code → Add to task
- Can be decomposed? → Split into subtasks → Close original task
- Ready to implement now? → Move to Ready
- Otherwise → Add "will-do-later" tag and keep in Backlog
↓
Move to next task (repeat until all are done)
See the canonical definition in agkan/SKILL.md (Tag Priority section).
agkan-icebox to review and promote icebox tasks to backlog firstnpx claudepluginhub gendosu/gendosu-claude-plugins --plugin agkan-skillsGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.