Plan Kanban work

← Back to skills

1. Read the project instructions and relevant product code from the integration or feature worktree. Read existing task/spec metadata through the canonical controller; do not assume `.juno_task/plan.md` exists. 2. Produce one concise PDR covering the goal, current behavior, scope, exclusions, risks, dependencies, acceptance criteria, and focused tests. Draft it in a fresh external file; do not place it in the product tree or a task body. 3. Preflight `yy ledger --help` and `yy ledger artifact --help`. Capture the PDR as a local immutable `report` Artifact Record with task/request provenance, then verify its ID, digest, size, retention, retrieval, and history. If the artifact API is unavailable, stop with the external draft intact and request an upgrade; never fall back to product `docs/`, task bodies/responses, new `.juno_task/specs`, or direct store edits. 4. Split only when pieces can be implemented and validated independently. Concurrent tasks must have explicit path ownership and dependencies. 5. Create tasks through routed `yy ledger` commands. Put concise durable requirements and acceptance criteria in each task body, record the PDR artifact ID in supported task fields/provenance, and relate follow-ups instead of reopening archived IDs. 6. Product documentation is only documentation shipped with the product. Never create controller-private tasks, ledger, state, artifacts, objects, specs, or receipts inside a product or feature worktree. 7. Do not start implementation, create worktrees, push, deploy, or mutate production unless the user separately asks.

Category: Business & Product
Repo: antigravity-awesome-skills
Path: skills/plan-ledger-tasks-yylo/SKILL.md
Updated: 10/11/2026, 8:00:18 PM

AI Summary

1. Read the project instructions and relevant product code from the integration or feature worktree. Read existing task/spec metadata through the canonical controller; do not assume `.juno_task/plan.md` exists. 2. Produce one concise PDR covering the goal, current behavior, scope, exclusions, risks, dependencies, acceptance criteria, and focused tests. Draft it in a fresh external file; do not place it in the product tree or a task body. 3. Preflight `yy ledger --help` and `yy ledger artifact --help`. Capture the PDR as a local immutable `report` Artifact Record with task/request provenance, then verify its ID, digest, size, retention, retrieval, and history. If the artifact API is unavailable, stop with the external draft intact and request an upgrade; never fall back to product `docs/`, task bodies/responses, new `.juno_task/specs`, or direct store edits. 4. Split only when pieces can be implemented and validated independently. Concurrent tasks must have explicit path ownership and dependencies. 5. Create tasks through routed `yy ledger` commands. Put concise durable requirements and acceptance criteria in each task body, record the PDR artifact ID in supported task fields/provenance, and relate follow-ups instead of reopening archived IDs. 6. Product documentation is only documentation shipped with the product. Never create controller-private tasks, ledger, state, artifacts, objects, specs, or receipts inside a product or feature worktree. 7. Do not start implementation, create worktrees, push, deploy, or mutate production unless the user separately asks. It is useful for product planning, requirements, operations, business workflows, and stakeholder communication. Source: antigravity-awesome-skills (skills/plan-ledger-tasks-yylo/SKILL.md).

Plan Kanban work

  1. Read the project instructions and relevant product code from the integration or feature worktree. Read existing task/spec metadata through the canonical controller; do not assume .juno_task/plan.md exists.
  2. Produce one concise PDR covering the goal, current behavior, scope, exclusions, risks, dependencies, acceptance criteria, and focused tests. Draft it in a fresh external file; do not place it in the product tree or a task body.
  3. Preflight yy ledger --help and yy ledger artifact --help. Capture the PDR as a local immutable report Artifact Record with task/request provenance, then verify its ID, digest, size, retention, retrieval, and history. If the artifact API is unavailable, stop with the external draft intact and request an upgrade; never fall back to product docs/, task bodies/responses, new .juno_task/specs, or direct store edits.
  4. Split only when pieces can be implemented and validated independently. Concurrent tasks must have explicit path ownership and dependencies.
  5. Create tasks through routed yy ledger commands. Put concise durable requirements and acceptance criteria in each task body, record the PDR artifact ID in supported task fields/provenance, and relate follow-ups instead of reopening archived IDs.
  6. Product documentation is only documentation shipped with the product. Never create controller-private tasks, ledger, state, artifacts, objects, specs, or receipts inside a product or feature worktree.
  7. Do not start implementation, create worktrees, push, deploy, or mutate production unless the user separately asks.

Use --id, not legacy --ID, for Kanban mutations. Return the task IDs and a short dependency/order summary.

$ARGUMENTS

When to Use

  • The user explicitly asks to plan or register work in the YYLO Ledger.
  • You need a concise Product Development Requirement (PDR) plus implementation-sized Ledger tasks with dependencies.

Limitations

  • Planning only: never start implementation, create worktrees, push, deploy, or mutate production from this skill.
  • Requires the installed yy ledger artifact API; if unavailable, stop with the external PDR draft intact - never fall back to product docs/, task bodies, or direct store edits.
  • Concurrent tasks need explicit path ownership and dependencies; relate follow-ups instead of reopening archived IDs.

Example

yy ledger --help
yy ledger artifact --help

Adapted from yylo-dev/yylo-skills (MIT) - v2.0.1; frontmatter, When to Use/Limitations, and safety boundaries added for upstream compliance.

Related skills