ShellGPT
Turns a plain-English request into the terminal command you need, and can run it for you
by Farkhod Sadykov
Plain-Language Session Report
by Pavlo KharmanskyiSkillWrite code
Tells you in one screen whether your AI agent finished the job and what it needs from you
The problem it solves
Your agent says it finished, but the summary is commit hashes and jargon, and you cannot tell whether you are done.
For People building with AI agentsReaders who do not code
- Before
- You read a wall of engineering detail and still have to ask whether anything is left for you.
- After
- You get about ten lines: what changed, bad news in its own row, and a verdict on whether it is done and what you need to do.
- Stars
- ★ 1.3k
- License
- MIT
- Cost
- Free
- Level
- Beginner
- Contents
- Instructions only
Overview
This skill from the Open Steps pack makes a coding agent end its work with a short report in plain words. You get a one- or two-sentence lead, a few rows for what was done, what went wrong and what was put off, and a verdict: fully done, anything needed from you, new unfinished work, and whether you can close the task. Each "yes" names its proof, unchecked items say "not checked", and the report is saved for the next session. It suits people who build products with an AI agent but do not read code.
It turns commit hashes and jargon into a ten-line verdict, and bad news gets its own row instead of hiding in the middle of a paragraph.
Try these first
Start with the first one: it is a small task you can check easily.
- Are we done? Give me the report and name the proof for every yes.First try · saves a report file
- What happened in this session? Keep it to one screen.
- What's the status of [ticket name]?
- Rewrite your last update as a done-or-not report in [language].
Example prompts written by Agent Flows Hub.
What you get
From the creator's documentation.
| Output | What it is | Reads |
|---|---|---|
| Lead | One or two sentences on what changed for the product, leading with the main result | This session |
| Checkmark rows | Two to five rows for what was done, what went wrong and what was put off | This session |
| Verdict | Fully done, anything needed from you, new unfinished work, and whether you can close | This session |
| Proof for each yes | Reads git and, if a pull request exists, its GitHub checks; unconfirmed items say not checked | Your repo and GitHub |
| Saved report | Kept in ~/.claude/open-steps/reports/ for the next session, never in your project | Your computer |
Good to know
At the end of a coding session, get a one-screen report that says whether the work is finished and what you need to do. The skill saves each report in a folder outside your project so the next session can start from it.
Limits
- It only uses proof it can gather in seconds and does not re-run your tests; anything it cannot confirm is marked "not checked".
- Its last step updates a BIG-PICTURE.md file through os-big-picture, a separate Open Steps skill, when that file exists or you have asked for one.
Notes by Agent Flows Hub (original project name: os-done-or-not), written from the creator's documentation, not from running it ourselves. os-done-or-not SKILL.md · Open Steps README
Stuck setting it up?
Tell us what happened. We take a look first; if it needs paid work, you get an estimate before anything starts.
More to write code
Other skills
- Planning with Files Makes your AI agent keep its plan, notes and progress in files so long tasks stay on track
- PM Resume Review Gives point-by-point feedback on a product manager resume, with suggested rewrites
- PPT Master Turns a document or a topic into a PowerPoint deck you can keep editing
- PRD Writer Writes a short product requirements document from your idea, users and success measure
- Product Deep-Dive Report Researches a product, company or idea online and writes a long PDF report in Chinese
- Product Listing Writer Writes titles, bullet points, descriptions and keywords for your online shop listings