Solo to Team Skills plugin
Four skills for a person with several AI assistants: a task register, evidence-based acceptance, beginner instructions
- 828 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- Instructions and a template only: the pack runs nothing and sets no timers
- The author has not verified the plugin structure for Claude Code
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/solo-to-team-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- Instructions and a template only: the pack runs nothing and sets no timers
- The author has not verified the plugin structure for Claude Code
This is third-party code. Review the repository files before installing.
What it does
Solo to Team is a set of four skills, a template and redacted cases for a person leading several long-lived AI assistants. The `dispatch-accept-verify` skill walks a task along a chain: before work is handed out, a "pending" line goes into the register with the assignee, the collector and the check time, then "taken", "result received" and "accepted" are recorded separately. `evidence-before-done` requires valid evidence for the current version of a result before anyone says "done": open it, recompute it, walk the steps. `beginner-ops-writing` teaches instructions a newcomer can follow in one go: one step at a time, unambiguous names and a description of what will appear on screen. Steps that only the owner can do (login, payment, confirmation) are marked "owner does this". `ops-case-notes` holds two redacted write-ups of real failures. The pack sets no timers and wakes no assistants (the plugin manifest has `hooks: {}`): time-based checks are done by a person or the agent host. It was verified in several hosts, with a marketplace manifest for Codex and copying into `~/.cursor/skills/` for Cursor. The skill texts are written in Chinese, and the English README is complete.
Who it is for. Founders and team leads running several AI assistants who lose track of work: unclear who took it, who will check and what counts as done.
Good fit when
- Work is handed to several assistants and you need a register showing who took what and when to check
- An assistant reports done and the result must be accepted only on evidence for the current version
- An instruction for a person with no experience has to work on the first try
- You are starting to lead an AI team and want to rehearse the cycle on a notebook table before connecting agents
- Before analyzing your own failure you want examples of hand-off and acceptance mistakes that already happened
Not a fit when
- You need timers and automatic wake-ups: the pack has none and they depend on the host
- You need a ready task scheduler or tracker: the pack gives a method and a register line template
- Chinese-language skills are inconvenient: the `SKILL.md` texts are in Chinese, although you can ask the assistant to reply in English
Example request
Use the dispatch-accept-verify skill: log a pending line for the task of writing meeting notes and check the result with evidence-before-doneLimitations
The documentation is a draft of version 0.9 and the author warns of changes before the public release. The `SKILL.md` texts and the register template are in Chinese, while the README and the worked example have English versions. The plugin structure is ready for Claude Code, but the author has not verified it there and left it out of launch support. The `dispatch-accept-verify` and `evidence-before-done` skills install together, since acceptance in the first relies on the second. The "about ten minutes" figure is a design goal, and only the Cursor CLI install was measured, at about 30 seconds.
How to disable. In Codex `codex plugin remove solo-to-team-skills` removes the plugin. In Cursor delete the copied folders from `~/.cursor/skills/`. The task register files you kept are yours and are deleted by hand.
FAQ
How do I install Solo to Team in Codex?
From the pack root run `codex plugin marketplace add "$(pwd)"`, then `codex plugin add solo-to-team-skills@solo-to-team-skills` and open a new session. The manifest is in `.agents/plugins/marketplace.json`.
Which skills should I enable first?
`dispatch-accept-verify` and `evidence-before-done`: acceptance in the first depends on evidence checking in the second. The other two, `beginner-ops-writing` and `ops-case-notes`, are optional.
What goes into the register?
One line per event: pending, taken, result received, accepted, sent back, cancelled, extended, escalated. The template is in `templates/task-register-line.md`.
Do I need Claude Code?
No. The author verified Codex, Cursor and other hosts, and prepared a Claude Code structure in `.claude-plugin/` without separate verification.
Does the pack remind me about checks?
No. The manifest has `hooks: {}`, so you or the agent host do the time-based checks, if the host has timers.
Discussion
No questions or reviews yet.
Related
Pluginscafe3310's personal skill collection: spec-driven project work, reviews, reports, browser, 3D and video in one repository
Eleven skills to stay aligned with an agent: plain-language restatement, recaps, a status board and a pull back to the goal.
18 portable skills for work where AI is weakest: messy voice input, message tone, small trend tests and safe shipping.
Third Brain Skills: a knowledge base and verification for agents
Third Brain Skills
21 skills build an Obsidian-style knowledge base with provenance, daily loops and verification before any done claim
293 skills in 16 categories for Codex, Claude Code and Cursor with an npx installer, provenance and license checks
A skill that cross-examines a plan before work starts: numbered questions in rounds with a recommended answer and no changes until you confirm
README in short
The page explains the idea: one person leads a team of AI assistants along the chain of hand out, collect, verify, decide. A table then separates what the skills do, what the host does and what stays with a human, with paths for a newcomer, a user of one assistant and an owner of several, and a main example with a fictional meeting. The end lists supported hosts with install commands and the measured time.