# AI use cases for Open Steps (kharmanskyi/open-steps) Repo: [kharmanskyi/open-steps](https://github.com/kharmanskyi/open-steps) | Use cases: 4 | Page: https://dedupedai.com/use-cases/kharmanskyi-open-steps/ A set of plain-language agent skills for Claude Code, Codex CLI, Cursor CLI and Gemini CLI that turn agent work into honest reports, clear verdicts and steps a non-engineer can follow. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Get a one-screen done-or-not report after each session: https://dedupedai.com/use-cases/kharmanskyi-open-steps/get-a-one-screen-done-or-not-report/ - Turn agent requests into numbered steps you can follow: https://dedupedai.com/use-cases/kharmanskyi-open-steps/turn-agent-requests-into-numbered-steps/ - Check another agent's "done" claims before accepting them: https://dedupedai.com/use-cases/kharmanskyi-open-steps/check-another-sessions-claims-before-accepting/ - Keep a plain-language status file for the whole project: https://dedupedai.com/use-cases/kharmanskyi-open-steps/keep-a-plain-language-project-status-file/ ## How to get a one-screen done-or-not report after each session with Open Steps Repo: [kharmanskyi/open-steps](https://github.com/kharmanskyi/open-steps) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Ask "are we done?" and get a verdict on done, needed input, new debt and whether it is safe to close. **Who it is for:** A founder or product owner building with an AI agent who cannot tell from commit hashes whether a task is finished. **The problem:** Agent reports are written in commit hashes and jargon, so the person directing the work cannot judge whether it is done or what they still owe. **The outcome:** A short report with a verdict row, a row for anything needed from you, a row for new debt, and a row for whether it is safe to close. ### Steps 1. Run `git clone https://github.com/kharmanskyi/open-steps.git` 2. Run `claude plugin marketplace add ./open-steps && claude plugin install open-steps@open-steps` (Claude Code) 3. Run `grep -q 'os-done-or-not' ~/.claude/CLAUDE.md 2>/dev/null || cat open-steps/docs/routing-block.md >> ~/.claude/CLAUDE.md` to add the routing block 4. Finish a unit of agent work and ask the agent "are we done?" 5. Read the verdict row first, then the row for anything needed from you ### Prompt to paste into your agent ```text In the open-steps repo at , use the os-done-or-not skill to report on the work just finished for . Expected output: a one-screen report with a Fully done row, an Anything needed from you row, a New debt row and a Safe to close row. Verify that each row states a verdict and does not rely on commit hashes alone. ``` ### Caveat The routing block must be added by hand on every tool, because the agent decides by itself whether to use a skill and the block is what tells it when. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/kharmanskyi-open-steps/get-a-one-screen-done-or-not-report/ ## How to turn agent requests into numbered steps you can follow with Open Steps Repo: [kharmanskyi/open-steps](https://github.com/kharmanskyi/open-steps) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README When the agent needs you to run, paste, click or approve something, get exact numbered steps instead. **Who it is for:** A non-technical operator who has to run commands or click through screens for an AI agent that is working on a project. **The problem:** When an agent needs a human action, the instructions are often vague or assume engineering knowledge, so the operator stalls or makes mistakes. **The outcome:** Numbered steps that a non-technical person can follow, with the agent trying everything it can itself first and asking only for what truly needs the person. ### Steps 1. Run `git clone https://github.com/kharmanskyi/open-steps.git` 2. Run `mkdir -p ~/.agents/skills && cp -R open-steps/skills/os-* ~/.agents/skills/` on Codex CLI, Cursor CLI or Gemini CLI 3. Add the routing block from `open-steps/docs/routing-block.md` to your tool's standing instructions file 4. Run `bash open-steps/doctor.sh` to check that the install is wired (inferred, check the README) 5. When the agent asks you to act, check that the answer uses numbered steps ### Prompt to paste into your agent ```text In the open-steps repo at , use the os-step-by-step skill for the next action that needs a human on . Expected output: numbered steps a non-technical person can follow, with only the actions that truly need the person listed. Verify that the agent tried every step it could do itself before asking. ``` ### Caveat The README notes that the Codex CLI, Cursor CLI and Gemini CLI setup has not been tested for Claude-only permission features, so expect the ask-first behavior on those tools to differ. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/kharmanskyi-open-steps/turn-agent-requests-into-numbered-steps/ ## How to check another agent's "done" claims before accepting them with Open Steps Repo: [kharmanskyi/open-steps](https://github.com/kharmanskyi/open-steps) | Difficulty: Advanced | Time: 45 minutes | Basis: From the README Verify each claim from another session against what actually happened, then get a next action. **Who it is for:** A developer or product lead who gets a "finished" message from another agent session and wants proof before merging or moving on. **The problem:** An agent that says it is done may have overstated its results, and checking every claim by hand takes time the person does not have. **The outcome:** Each claim is checked against what actually happened, with a plain statement of what to do about any gaps. ### Steps 1. Run `git clone https://github.com/kharmanskyi/open-steps.git` 2. Install the skills for your tool (Claude Code: `claude plugin marketplace add ./open-steps && claude plugin install open-steps@open-steps`) 3. Ask the agent to check the other session's pull request with `gh pr view` and `gh pr checks` 4. Read the list of checked claims and the next action it recommends ### Prompt to paste into your agent ```text In the open-steps repo at , use the os-check-work skill to verify the claims in the report from about pull request . Expected output: each claim marked against what actually happened, followed by one recommended action. Verify that no claim is accepted without a check. ``` ### Caveat Pull request reads depend on gh being installed and signed in, and without it the output says "not checked" rather than verifying. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/kharmanskyi-open-steps/check-another-sessions-claims-before-accepting/ ## How to keep a plain-language status file for the whole project with Open Steps Repo: [kharmanskyi/open-steps](https://github.com/kharmanskyi/open-steps) | Difficulty: Advanced | Time: 45 minutes | Basis: From the README Maintain one BIG-PICTURE.md with every feature, its progress, unused parts and queued work. **Who it is for:** A product owner running several features with an agent who wants one standing view of where the project stands. **The problem:** Project status lives in scattered commits and session logs, so the owner cannot see the whole product or what is queued without reading code. **The outcome:** A BIG-PICTURE.md file in plain words that lists each feature and how far it got, unused parts, and the queue, with optional tickets opened in a tracker after your approval. ### Steps 1. Run `git clone https://github.com/kharmanskyi/open-steps.git` 2. Install the skills for your tool, then add the routing block 3. Ask the agent where the project stands, so os-big-picture writes or updates BIG-PICTURE.md 4. Review the file and reply yes or no to opening queued items as tickets ### Prompt to paste into your agent ```text In the open-steps repo at , use the os-big-picture skill to update BIG-PICTURE.md for the project at . Expected output: a plain-language file listing each feature with its progress, any unused parts and the queued work. Verify that no task was invented and nothing was deleted. ``` ### Caveat The README says the file stays out of your commits unless you add it yourself, so you must commit it deliberately if you want it shared. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/kharmanskyi-open-steps/keep-a-plain-language-project-status-file/