aei-soli/claude-skills — explained in plain English
Analysis updated 2026-05-18
Keep a persistent tracker of every project you're running with AI, including priority and next steps.
Turn your project tracker into a one week sprint plan that assigns tasks to the right AI model.
Avoid running out of premium AI usage mid-week by scheduling around budget reset cycles.
Get daily check-ins and a remediation plan when your week's plan starts slipping.
| aei-soli/claude-skills | 0xkinno/neuralvault | 0xlocker/d17-contracts | |
|---|---|---|---|
| Stars | 1 | 1 | 1 |
| Language | — | TypeScript | Solidity |
| Setup difficulty | easy | hard | hard |
| Complexity | 1/5 | 4/5 | 5/5 |
| Audience | pm founder | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
This project is a Claude skill and plugin called the AI Portfolio Manager, built around what the author calls the SprintFleet Method. It is meant for people juggling many projects at once with AI help, who tend to run into the same three problems: their premium AI usage budget runs out partway through the week, tasks that need a human to step in pile up and get forgotten, and projects stall because nothing keeps track of what the next step was supposed to be. The method works in two layers. The first is a portfolio layer, a persistent tracker listing every project, its priority, its status, and one clear next step, along with roadmap sheets for bigger projects where each task comes with a ready to paste AI prompt, so any future session can pick it up without needing to re-explain context. The second is a sprint layer, which turns that tracker into a one week plan, assigning each task to the right resource, whether that is a premium AI model, a cheaper workhorse model, a research model, an overnight batch job, or a human, scheduled around when usage budgets actually reset. Each day gets exactly one starred task that a human must handle, since the README notes that human gated work is what tends to slip. There are four planning styles to choose from: quality first, balanced, budget first, and deadline first. The README describes internal benchmark testing where plans made with the skill passed more quality checks than plans made without it. It installs as a Claude plugin through the desktop app, Cowork, or Claude Code, and is released under the MIT license. A paid edition with more automation is planned for the future.
A Claude plugin that helps you manage many AI-assisted projects at once by tracking next steps and scheduling weekly tasks around your usage budget.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly pm founder.
This repo across BitVibe Labs
Verify against the repo before relying on details.