# AI use cases for Sure Forge (Da7-Tech/SureForge) Repo: [Da7-Tech/SureForge](https://github.com/Da7-Tech/SureForge) | Use cases: 4 | Page: https://dedupedai.com/use-cases/da7-tech-sureforge/ SureForge is a plain-text, instruction-only Agent Skill that tells an AI agent to research, ask, plan, verify, and get independent review before it calls complex work done. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Plan a multi-step change before any code is written: https://dedupedai.com/use-cases/da7-tech-sureforge/plan-multi-step-change-before-building/ - Run a high-stakes change in full mode with extra checks: https://dedupedai.com/use-cases/da7-tech-sureforge/run-high-stakes-work-in-full-mode/ - Check that an installed Sure Forge copy matches the source: https://dedupedai.com/use-cases/da7-tech-sureforge/check-installed-skill-matches-source/ - Get a delivery report of what was checked and what was not: https://dedupedai.com/use-cases/da7-tech-sureforge/get-a-delivery-report-of-what-was-checked/ ## How to plan a multi-step change before any code is written with Sure Forge Repo: [Da7-Tech/SureForge](https://github.com/Da7-Tech/SureForge) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Make an agent research, ask, and plan a big change before it builds anything. **Who it is for:** A developer about to hand a substantial multi-step change to an AI agent who wants to check the plan first. **The problem:** Agents often start building before the request is understood, and they treat skipped questions as a yes. **The outcome:** A plan that maps each acceptance criterion to a step and a way to verify it, reviewed before any build starts. ### Steps 1. Run `npx skills add Da7-Tech/SureForge` from your project. 2. Ask your agent: Use Sure Forge for this task. Research the important unknowns before you ask me anything, then give me a plan I can check before you build. 3. Answer the questions the agent asks, and check that it does not treat any skipped question as an answer. 4. Read the plan gate result, which is READY, REPAIR, or BLOCKED. 5. Let the agent start the execute phase only after the plan gate is READY. ### Prompt to paste into your agent ```text Using the Da7-Tech/SureForge skill, research the unknowns in before asking me questions, then produce a plan that maps each acceptance criterion to a step and a verification method. Expected output: the plan and a plan gate result of READY, REPAIR, or BLOCKED. Verify that no build step appears before the plan gate is READY. ``` ### Caveat The README gives the skill's wording and gates, but it does not claim measured gains on real underspecified work. 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/da7-tech-sureforge/plan-multi-step-change-before-building/ ## How to run a high-stakes change in full mode with extra checks with Sure Forge Repo: [Da7-Tech/SureForge](https://github.com/Da7-Tech/SureForge) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Use the skill's full mode for risky or hard-to-reverse work with more verification. **Who it is for:** An engineer or team lead preparing a risky or hard-to-reverse change who wants independent review at every gate. **The problem:** For high-consequence work, a single agent self-check is not enough, and a missing reviewer can be quietly skipped. **The outcome:** A run where each gate has agent and reviewer verification, and any missing reviewer is reported instead of faked. ### Steps 1. Install the skill for your agents with `npx skills add Da7-Tech/SureForge --agent claude-code --agent codex --agent cursor --agent devin --agent hermes-agent -y`. 2. Ask your agent: Use Sure Forge in full mode. Do not pass a gate without three verification methods of your own and three from an independent reviewer. If a reviewer or tool is missing, tell me instead of pretending. 3. Confirm the agent declares the tier it is using before it starts. 4. If no independent reviewer is available, accept that the agent will report this, and decide whether to proceed. 5. Read the coverage ledger and evidence record the agent leaves behind. ### Prompt to paste into your agent ```text Using the Da7-Tech/SureForge skill in full mode, plan and carry out in . Expected output: a gate result for each phase, the tier used, and a coverage ledger listing what was verified and what was not. Verify that the agent states plainly if no independent reviewer was available. ``` ### Caveat Full mode depends on an independent reviewer being available, and the README says the skill reports it when one is missing. 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/da7-tech-sureforge/run-high-stakes-work-in-full-mode/ ## How to check that an installed Sure Forge copy matches the source with Sure Forge Repo: [Da7-Tech/SureForge](https://github.com/Da7-Tech/SureForge) | Difficulty: Advanced | Time: 30 minutes | Basis: From the README Confirm that an installed copy of the skill is byte-identical to the repository version. **Who it is for:** A maintainer or team lead who installed Sure Forge into a shared project and wants to confirm the copy has not drifted. **The problem:** A copied skill folder can be incomplete or edited, and an agent following a changed copy will behave differently. **The outcome:** A verification result that shows whether the installed folder matches the source, plus unit test results. ### Steps 1. Clone the Da7-Tech/SureForge repository and open its root folder. (inferred, check the README) 2. Run `python3 -B -m scripts.verify_install --installed path/to/installed/sureforge` with the path to your installed copy. 3. Run `python3 -B -m unittest discover -s tests -v` to check the package tests. 4. Note any file the install check reports as different, and reinstall or recopy that file. 5. Rerun the install check until it passes. ### Prompt to paste into your agent ```text In the Da7-Tech/SureForge repository, run the install verification against and then the unit tests. Expected output: a pass or a list of mismatched files, and the test summary. Verify that the commands ran from the repository root with Python 3.11 or newer. ``` ### Caveat The README lists these checks, but it does not say how a Python environment should be set up beyond Python 3.11 or newer. 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/da7-tech-sureforge/check-installed-skill-matches-source/ ## How to get a delivery report of what was checked and what was not with Sure Forge Repo: [Da7-Tech/SureForge](https://github.com/Da7-Tech/SureForge) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Receive agent-written work with a clear list of verified, reused, and unchecked items. **Who it is for:** A person receiving work from an AI agent who needs to know what was actually tested before accepting it. **The problem:** Agents sometimes check a sample and call the work complete, which leaves gaps the receiver cannot see. **The outcome:** A delivery report that lists what was verified, what was reused, and what was not checked, with any blocked checks named. ### Steps 1. Run `npx skills add Da7-Tech/SureForge` from your project. 2. Ask your agent to finish the work with the deliver phase and to report what was verified, what was reused, and what was not checked. 3. Check that any check the agent could not run, such as a visual render, is reported as blocked and not claimed as done. 4. Review the evidence record and coverage ledger before accepting the work. 5. Send back any item marked unresolved or not checked for another round. ### Prompt to paste into your agent ```text Using the Da7-Tech/SureForge skill, freeze the candidate for , inspect every agreed unit on that exact version, and report what was verified, what was reused, and what was not checked. Expected output: a delivery gate result and a coverage ledger. Verify that every item in the report is backed by evidence or marked as not checked. ``` ### Caveat The README says the coverage ledger and evidence record come from the skill's templates, so their completeness depends on the agent following the text. 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/da7-tech-sureforge/get-a-delivery-report-of-what-was-checked/