threads
Skill1 starsChanged 23 days agoMITView on GitHub
ADHD-friendly branching task tracker. Use when the user dumps multiple or branching requests, switches tasks, says "bt" / "backtrack" / "switch to" / "where am i" / "done" / "snooze", or whenever work threads should be captured or resumed so nothing gets dropped. Auto-captures every actionable request into a persistent tree with a priority queue.
10 other projects use a version of this.
Adopted by
Most recently taken up by
- uplika/uplika-plugintheir own version, first seen Oct 25, 2026
- matthull/my-skillstheir own version, first seen Oct 20, 2026
- obra/muse-skillstheir own version, first seen Oct 13, 2026
- shabash-dev/shabash-pluginstheir own version, first seen Oct 13, 2026
- AceDataCloud/Skillstheir own version, first seen Oct 6, 2026
- thenavidm/threads-mcp-clitheir own version, first seen Oct 5, 2026
Other versions10 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 10
- obra/muse-skills★ 20same name
- AceDataCloud/Skills★ 17same name
- jacobwell/muse-skills★ 8same name
- thenavidm/threads-mcp-cli★ 4same name
- uplika/uplika-plugin★ 3same name
- retrodigio/claude-channel-slack★ 2same name
- andrmaz/spec-driven-architecture★ 1same name
- Dimks777/aiclub★ 0same name
- matthull/my-skills★ 0same name
- shabash-dev/shabash-plugins★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository9 other files it gives its agents.
- bughuntSkill
- commitSkill
- comprehensive-reviewSkill
- councilSkill
- dequeueSkill
- enqueueSkill
- github-issueSkill
- github-prSkill
- md2pdfSkill
The file
We split this skill into the rules it gives an agent
3 rules. 0 are common, 2 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 9 edited versions we read keep these.
2
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 12
Run the CLI with node:
- Very specific: line 20
When the user states an actionable request, capture it.
- Distinctive: line 24
After capturing, keep doing what the user actually asked.
“Common” is measured: kept by at least 6 in 10 of the 9 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
threads
A persistent, branching task tree so neither you nor the user loses track of work. The user has ADHD and dumps branching ideas faster than they can file them; your job is to file them so nothing is dropped. State lives at ~/.claude/threads.json. Drive it with the CLI, never hand-edit the JSON.
Per-chat by default: each chat is its own tree with its own current (scoped by CLAUDE_CODE_SESSION_ID), so chats never tangle. bare threads (or here) shows where you are plus one layer; all zooms out to every chat's full tree.
Run the CLI with node:
node ~/.claude/skills/threads/bin/threads.mjs <command>The discipline (no dropped balls)
- When the user states an actionable request,
captureit. Several requests in one message means several captures. - A request blurted while you are mid-task auto-nests as a child of the current task.
- A task leaves the tree only two ways:
done, or the user explicitly kills it. Never silently. doneon a parent drops it and shifts its children up to the grandparent, so open work is never left stranded under a finished branch (a done leaf stays as a✓). Runcompactto apply that sweep to an existing tree where parents were finished before their children.- After capturing, keep doing what the user actually asked. The rest wait in the queue, surfaced 1-3 at a time, never the whole backlog.
Commands
| Command | Does |
capture "<name>" | file a new task (auto-nests under current) |
switch "<query>" | resume an existing task by name |
bt | return to the parent task |
done ["<query>"] | complete a task; a done parent's children shift up to the grandparent |
snooze "<query>" [days] | defer a task; it resurfaces later |
compact | sweep finished branches: drop done parents + shift their children up (done leaves stay) |
here (or no arg) | where you are + one layer: breadcrumb, current, direct children (the ▸ marks you) |
all | every chat's full tree, grouped by project (zoom out) |
Triggers
- "bt", "backtrack", "go back" ->
bt - "switch to X", "back to X" ->
switch "X" - "done", "finished", "that's fixed" ->
done - "snooze X", "not now", "later" ->
snooze "X" - "where am i", "show tree", "what's open" ->
here - the user dumps one or more new tasks ->
captureeach
Glyphs (emoji-free, single-width so columns align)
▸ here · ● active · ○ paused · ✓ done · ✕ blocked · ◦ snoozed
Keep upkeep invisible
The user will not maintain this manually. Capture silently as part of doing the work; do not narrate the bookkeeping. Surface the tree only when asked or when resuming.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.