Skip to content

mxyhi/ok-skills / ledger-tasks-yylo

ledger-tasks-yylo

Verified

Durable in-repo task state for coding agents — Kanban tasks with stable IDs, evidence-backed status marks, blocked-by dependencies, and next-ready selection. Use when starting, selecting, sequencing, or closing implementation tasks for a coding-agent project.

YYLO Ledger Task Management

Use this skill when a coding agent needs durable, reviewable task state that lives in the repository instead of a chat transcript. YYLO Ledger is the source of truth for task state; this skill teaches the working loop.

Adapted from the canonical ledger-tasks-yylo skill in yylo-dev/yylo-skills; see the upstream SKILL.md for the full CLI reference (archives, multi-directory merges, controller routing).

When to Use

  • Decide which task is safe to start next (ready = all blockers done)
  • Start assigned work and record an evidence receipt (mark in_progress --response)
  • Close a task with what changed and how it was tested (mark done --response --commit)
  • Declare dependencies between tasks before creating them
  • Sequence parallel work topologically (order --scores)

Workflow

  1. Pick work: yy ledger ready (or yy ledger list --status todo,in_progress).
  2. Inspect full state before any mutation: yy ledger get TASK_ID (includes resolved blocked_by and related tasks).
  3. Start: yy ledger mark in_progress --id TASK_ID --response "Starting: <plan>".
  4. Finish: yy ledger mark done --id TASK_ID --response "Done: <what> — tested <how>" --commit <sha>.
  5. If the task cannot proceed, keep the blocker explicit instead of closing it silently.

Core Commands

yy ledger create "Short, one-iteration task" --status backlog --tags feature,backend
yy ledger list --status todo,in_progress --limit 10
yy ledger search --tag backend --open
yy ledger get TASK_ID
yy ledger update TASK_ID --tags backend,urgent
yy ledger deps add --id TASK_ID --blocked-by BLOCKER1 BLOCKER2
yy ledger ready --tag backend
yy ledger order --scores
yy ledger archive TASK_ID        # soft delete, preserves data

Inline body markup is parsed on create/update, so dependencies can be declared in prose:

[blocked_by]TASK_ID[/blocked_by]     — this task waits on TASK_ID
[task_id]RELATED_ID[/task_id]        — cross-reference a related task

Best Practices

  • Keep tasks small enough to finish in one iteration without filling the context window.
  • Status flow: backlog -> todo -> in_progress -> done (archive for abandoned tasks).
  • Always pass --response on mark; attach --commit when marking done so task history links to the diff.
  • Use ready before starting, and order --scores when planning parallel pipelines.
  • Read state with get before mutating; never bypass the CLI with direct file edits.

Prerequisites

  • YYLO CLI: npm install --global '@yylo/cli@latest' (Node.js 20.10+, npm, Git). yy ledger delegates to the standalone YYLO Ledger Git-native task store (MIT).
  • Task state lives under the repository's .juno_task/ directory and is committed with the project, so any agent or reviewer can reconstruct decisions from history.
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall