From lo
Captures ideas and rich conversation context for later. Creates a backlog entry linked to a detailed conversation summary in .lo/park/. Use when user says "park this", "save this for later", "park", "remember this idea", or "/lo:park". Not for planning or execution — use /lo:plan to design, /lo:work to build.
How this skill is triggered — by the user, by Claude, or both
Slash command
/lo:lo-parkThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Captures ideas and conversation context for later. When you've been discussing something with Claude and want to save the thinking without committing to planning or building, park it.
Captures ideas and conversation context for later. When you've been discussing something with Claude and want to save the thinking without committing to planning or building, park it.
/lo:park/lo:plan/lo:plan (creates backlog entry + plans in one motion)/lo:work (creates backlog entry + executes)/lo:ship.lo/ directory must exist. If missing, tell user to run /lo:setup first.last_feature and last_task from BACKLOG.md frontmatter for ID allocation.Detect mode from arguments:
/lo:park (no args) → Mode 1: Capture from Conversation/lo:park feature "name" → Mode 2: Quick Park (backlog entry only)/lo:park task "name" → Mode 2: Quick Park (backlog entry only)This is the primary mode. The user has been discussing something and wants to preserve the thinking.
Step 1: Determine what to capture.Suggest a name based on the conversation topic:
What should I call this?
1. [suggested name based on conversation]
2. Something else
Classify as feature (f{NNN}) or task (t{NNN}) based on scope. Features are bigger and need design. Tasks are smaller, just-do-it items. If scope is ambiguous, ask:
Is this a feature (bigger, needs design) or a task (smaller, just do it)?
Step 2: Generate the conversation capture.
Read the conversation context and produce a rich, near-verbatim summary that preserves the thinking. This is NOT a structured extraction — it reads like meeting notes that capture the actual flow of ideas.
Include all of these:
The capture should be long enough that future-you can read it and be immediately back in the headspace. Err on the side of too much context rather than too little.
Format the capture file per references/park-format.md:
# <id> — <name>
parked: <YYYY-MM-DD>
<Rich narrative capture. Multiple paragraphs.
Preserves the flow of thinking, not just the conclusions.>
Step 3: Present for review.
Show the complete capture to the user:
Here's what I captured. Review and approve, or tell me what to change:
---
[full capture content]
---
Save this? (yes / edit / redo)
HARD GATE: Do not write any files until the user approves. If the user says "edit" or "redo", ask what to change, regenerate, and re-present. Repeat until approved.
Step 4: Persist.
After the user approves:
# Read the counter from frontmatter
grep 'last_feature:' .lo/BACKLOG.md # → last_feature: 9
# Next feature ID = last_feature + 1 → f010
.lo/park/ directory if it doesn't exist:mkdir -p .lo/park
# Path: .lo/park/<id>-<slug>.md
# Example: .lo/park/f010-image-gen.md
# f010 — Image Generation for MDX Pipeline
parked: 2026-03-12
Started by exploring whether we could auto-generate hero images...
[rich narrative continues]
For features, add under ## Features:
- [ ] f010 Image Generation for MDX Pipeline
AI-powered image generation integrated into the article publishing pipeline.
[parked](.lo/park/f010-image-gen.md)
For tasks, add under ## Tasks:
- [ ] t016 Fix dark mode toggle
[parked](.lo/park/t016-dark-mode.md)
# Before:
last_feature: 9
# After:
last_feature: 10
Update the updated: date in BACKLOG.md frontmatter to today.
Report:
Parked: f010 "Image Generation for MDX Pipeline"
Capture: .lo/park/f010-image-gen.md
Pick this up later with /lo:plan f010 or /lo:work f010
Just a backlog entry. No capture file. Use when the user provides a name directly.
grep 'last_task:' .lo/BACKLOG.md # → last_task: 15
# Next task ID = 15 + 1 → t016
If feature and no description was provided, ask for 1-2 sentences describing it.
Add the entry to BACKLOG.md under the appropriate section. No [parked] link — this is a plain backlog item.
For features:
- [ ] f010 Real-time Collab Editing
Collaborative editing for field notes using CRDTs.
For tasks:
- [ ] t016 Fix dark mode toggle
updated: date to today.Parked: <id> "<name>"
.lo/ directory doesn't exist → Stop. Tell the user: "Run /lo:setup first to initialize the project."---
updated: <today>
last_feature: 0
last_task: 0
---
## Features
## Tasks
User: park this
Assistant suggests name: "Image Generation for MDX Pipeline" (feature). User confirms.
Assistant generates rich capture preserving the full conversation thinking, presents it for review.
User: yes
Parked: f009 "Image Generation for MDX Pipeline"
Capture: .lo/park/f009-image-gen.md
Pick this up later with /lo:plan f009 or /lo:work f009
User: /lo:park feature "real-time collab editing"
Assistant reads BACKLOG.md, gets next feature ID (f010), asks for a brief description since none was provided.
User: Collaborative editing for field notes using CRDTs.
Assistant adds backlog entry with description, increments counter.
Parked: f010 "Real-time Collab Editing"
User: /lo:park task "fix dark mode toggle"
Assistant reads BACKLOG.md, gets next task ID (t016), adds entry directly. No description prompt needed for tasks.
Parked: t016 "Fix dark mode toggle"
npx claudepluginhub looselyorganized/lo-plugin --plugin loQuickly dumps raw ideas, bugs, or feature requests into the backlog as unrefined tasks without asking questions or starting a conversation. Use when you want to offload a thought and keep moving.
Interactively adds features, ideas, or requirements to project backlog by prompting for details, creating docs/features/[id]/idea.md with YAML frontmatter, validating duplicates from DASHBOARD.md, and optionally staging git changes.
Captures a single user-voiced idea into a provenance-tracked folder in an ideas repo, with optional GitHub issue mirror. For recording one idea at a time, not bulk extraction.