A
aaroncohn/openspec-conductor-powerups
Opinionated Claude Code skills for OpenSpec inside Conductor: start a batch of parallel explorations, babysit…
View on GitHub ↗Link checked 2026-11-12 · automated check blocked (HTTP 429), status unconfirmed
- Star count
- 1 ★ · aaroncohn/openspec-conductor-powerups · read from the repository page · (checked 2026-11-13 20:00:50)
- Source
- read directly from its GitHub page · type: open-source project
- Harvest history
- first seen 2026-10-29 · last seen 2026-11-14 · seen 17× · listed by 1 sources
- Score
- score 103 · 1 ★ · 1 sources · repo
FAQ (answers come from our records)
Who is it for?
We filed it under these tags: Agent 技能. The tags come from the entry itself, not a taxonomy we imposed.
Who maintains it?
repository
aaroncohn/openspec-conductor-powerups;first catalogued 2026-10-29;last seen 2026-11-14。There is no official certification or quality rating here — we record sources and dates only.# openspec-conductor-powerups
以下摘录自该仓库的 README(片段) · 在 GitHub 读完整 README →
Opinionated Claude Code skills for people who build with OpenSpec inside Conductor.
Conductor lets you run many coding agents side by side, each in its own workspace. OpenSpec gives each piece of work a plan (explore → propose → apply → archive). These skills handle the parts in between: deciding what to start next, getting everything that's been proposed shipped, and cleaning up afterwards.
| Skill | What it does |
| new-batch-explore | Reads your roadmap, picks the next 2–5 milestones that can be built in parallel without colliding, and opens a Conductor workspace for each one, already in /opsx:explore. |
| babysitter | After you've explored and proposed in several workspaces, run this in one more workspace. It messages each workspace's Claude session, answers their questions for you, and keeps them moving until every change is tested, merged, archived, and released to production. If safe-to-archive is installed, each workspace runs it last. Then it walks you through the decisions it made, one at a time with a recommendation, and the things only you can test, one at a time with full steps and a shortcut (a prompt for your browser agent, or an offer to do most of it). |
| safe-to-archive | Checks whether a workspace can be archived without losing anything: uncommitted or unpushed work, open PRs, stashes, unarchived OpenSpec changes, files that only exist in this workspace, running processes. Read-only. |
| browser-handoff | For cloud workspaces: when an agent needs something done in your real browser (a dashboard setting, a secret pasted into a web UI), it writes one self-contained task block, with the exact account and your project's rules, for you to paste into a local claude --chrome session. |
| better-decision | Before a choice that matters, stops and lays out at least four real options, each with pros, cons, how hard it is to undo, and a rating, then recommends one and waits for your pick. In an OpenSpec change, it records the pick in design.md. Use it anytime, in any workspace. |
## The loop
/new-batch-explore → N new workspaces, each exploring one milestone
/opsx:explore, /opsx:propose in each workspace (that's you, thinking with Claude)
/babysitter → apply, test, merge, archive, release (hands off)
/safe-to-archive in each workspace (babysitter runs it for you) → archive it in Conductor
Each skill works in both local (Mac) and cloud Conductor workspaces. It checks CONDUCTORISLOCAL and, in the cloud, uses the conductor CLI in place of things that only exist on your Mac (the conductor:// link, Conductor's local database). The cloud steps are in each skill's cloud.md. See Local vs cloud for what differs.
## Install
In Claude Code:
/plugin marketplace add aaroncohn/openspec-conductor-powerups
/plugin install powerups@openspec-conductor-powerups
The skills are then /powerups:babysitter, /powerups:new-batch-explore, and so on.
Stay on the latest version. There's no version number: every commit to main is a release. Turn on auto-update once and Claude Code picks up new commits in the background after a session starts: in /plugin, open Marketplaces, pick openspec-conductor-powerups, and choose Enable auto-update. Or add it to ~/.claude/settings.json:
Conductor lets you run many coding agents side by side, each in its own workspace. OpenSpec gives each piece of work a plan (explore → propose → apply → archive). These skills handle the parts in between: deciding what to start next, getting everything that's been proposed shipped, and cleaning up afterwards.
| Skill | What it does |
| new-batch-explore | Reads your roadmap, picks the next 2–5 milestones that can be built in parallel without colliding, and opens a Conductor workspace for each one, already in /opsx:explore. |
| babysitter | After you've explored and proposed in several workspaces, run this in one more workspace. It messages each workspace's Claude session, answers their questions for you, and keeps them moving until every change is tested, merged, archived, and released to production. If safe-to-archive is installed, each workspace runs it last. Then it walks you through the decisions it made, one at a time with a recommendation, and the things only you can test, one at a time with full steps and a shortcut (a prompt for your browser agent, or an offer to do most of it). |
| safe-to-archive | Checks whether a workspace can be archived without losing anything: uncommitted or unpushed work, open PRs, stashes, unarchived OpenSpec changes, files that only exist in this workspace, running processes. Read-only. |
| browser-handoff | For cloud workspaces: when an agent needs something done in your real browser (a dashboard setting, a secret pasted into a web UI), it writes one self-contained task block, with the exact account and your project's rules, for you to paste into a local claude --chrome session. |
| better-decision | Before a choice that matters, stops and lays out at least four real options, each with pros, cons, how hard it is to undo, and a rating, then recommends one and waits for your pick. In an OpenSpec change, it records the pick in design.md. Use it anytime, in any workspace. |
## The loop
/new-batch-explore → N new workspaces, each exploring one milestone
/opsx:explore, /opsx:propose in each workspace (that's you, thinking with Claude)
/babysitter → apply, test, merge, archive, release (hands off)
/safe-to-archive in each workspace (babysitter runs it for you) → archive it in Conductor
Each skill works in both local (Mac) and cloud Conductor workspaces. It checks CONDUCTORISLOCAL and, in the cloud, uses the conductor CLI in place of things that only exist on your Mac (the conductor:// link, Conductor's local database). The cloud steps are in each skill's cloud.md. See Local vs cloud for what differs.
## Install
In Claude Code:
/plugin marketplace add aaroncohn/openspec-conductor-powerups
/plugin install powerups@openspec-conductor-powerups
The skills are then /powerups:babysitter, /powerups:new-batch-explore, and so on.
Stay on the latest version. There's no version number: every commit to main is a release. Turn on auto-update once and Claude Code picks up new commits in the background after a session starts: in /plugin, open Marketplaces, pick openspec-conductor-powerups, and choose Enable auto-update. Or add it to ~/.claude/settings.json:
Listed badge (optional)
Copy this line into the README if you want the listing mark. Optional, and it does not affect being listed.
[](https://anclum.com/en/s/aaroncohn-openspec-conductor-powerups/)
- Source
- github · https://github.com/aaroncohn/openspec-conductor-powerups
- Browse by field
- Agent skills
- Traction
- 1 ★
- added
- today
More like this
Filed under
Every tag opens the full list for that filter. Counts are the live catalogue figures.