# AI use cases for Forest (mikamikasuki/forest) Repo: [mikamikasuki/forest](https://github.com/mikamikasuki/forest) | Use cases: 4 | Page: https://dedupedai.com/use-cases/mikamikasuki-forest/ FOREST is an open-source research agent framework that takes a research question through literature review, hypotheses, real experiments, analysis, and a full manuscript, keeping the work in an editable research graph. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Run a research project from question to draft paper: https://dedupedai.com/use-cases/mikamikasuki-forest/run-full-research-project-in-web-workspace/ - Run a research project from the terminal without the Web build: https://dedupedai.com/use-cases/mikamikasuki-forest/drive-research-project-headless-from-cli/ - Clone a public GitHub repo and inspect its run diagnostics: https://dedupedai.com/use-cases/mikamikasuki-forest/run-experiments-from-public-github-repository/ - Verify the exact CSV bytes behind a paired statistics run: https://dedupedai.com/use-cases/mikamikasuki-forest/audit-paired-statistics-input-bytes/ ## How to run a research project from question to draft paper with Forest Repo: [mikamikasuki/forest](https://github.com/mikamikasuki/forest) | Difficulty: Intermediate | Time: 1 hour | Basis: From the README Start a FOREST project in the browser and let it move from research goal to experiments and manuscript. **Who it is for:** A researcher with an early question who wants a guided, visual workspace for planning and running experiments with a set budget. **The problem:** Keeping ideas, experiments, results, and paper sections connected across weeks of work is hard in a long chat history. **The outcome:** A project with a research goal, a budget, an editable research graph, and saved runs, logs, metrics, and artifacts that you can inspect and redirect. ### Steps 1. Clone the repository with `git clone https://github.com/mikamikasuki/FOREST.git` and enter it with `cd FOREST`. 2. Install dependencies with `python3.11 scripts/install.py`. 3. Start FOREST with `.venv/bin/forest serve`. 4. Open http://127.0.0.1:8000/, configure a model provider in Settings, and create a project with a research goal and an explicit spending budget. 5. Choose a research mode (Manual, Assisted, or Auto) and review the research graph before letting tasks execute. ### Prompt to paste into your agent ```text Using the FOREST repository at , install the Web workspace and start it with the project's serve command. Create a project with the research goal and a spending budget of , then report the project ID and the first graph nodes that were created. Verify that the budget and goal appear in the project settings before any task runs. ``` ### Caveat Local execution runs generated code with the worker account's permissions, and a separate working directory does not isolate it at the operating-system level. Risk, credentials: Model provider credentials must be kept outside the repository and entered only in the local Settings configuration. 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/mikamikasuki-forest/run-full-research-project-in-web-workspace/ ## How to run a research project from the terminal without the Web build with Forest Repo: [mikamikasuki/forest](https://github.com/mikamikasuki/forest) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Use the FOREST CLI and API service to create and monitor a research project on a server or in scripts. **Who it is for:** An engineer or lab member working on a remote machine who wants to manage research projects from the command line or from an automated script. **The problem:** A Web browser is not available on the compute host, and the research state needs to be controlled from scripts. **The outcome:** A running headless FOREST service with a project initialized from the CLI and a status report showing its current state. ### Steps 1. Install without Node.js or the Web build with `python3.11 scripts/install.py --headless`. 2. Start the API and worker with `.venv/bin/forest serve --headless --data-dir ./var` and keep that terminal running. 3. In a second terminal, create a project with `.venv/bin/forest init --name "My research" --goal "Your research question"`. 4. Check progress with `.venv/bin/forest status` and list commands with `.venv/bin/forest --help`. ### Prompt to paste into your agent ```text In the FOREST repository at , run the headless install, start the service with a data directory of , and initialize a project with the goal . Report the output of the status command. Verify that the project name in the status output matches the name you passed to init. ``` ### Caveat The CLI guide and the numerical experiment example are linked from the README, and the example runs without model calls, so use it to check the setup before adding a provider. 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/mikamikasuki-forest/drive-research-project-headless-from-cli/ ## How to clone a public GitHub repo and inspect its run diagnostics with Forest Repo: [mikamikasuki/forest](https://github.com/mikamikasuki/forest) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Bring a public GitHub codebase into a FOREST project and check the source commit and run diagnostics. **Who it is for:** A researcher whose experiments depend on an existing public GitHub codebase and who needs a traceable source commit for each run. **The problem:** It is unclear which commit of a research codebase produced a result, and failed runs are hard to diagnose. **The outcome:** A repository input in the selected project with a run ID, its recorded source commit, and execution diagnostics. ### Steps 1. Make sure Git is installed on the worker, and that OpenSSH is available if you use SSH. 2. Clone the repository into the selected project with `.venv/bin/forest repo clone https://github.com/OWNER/REPOSITORY.git --wait`. 3. Note the returned run ID and inspect its source commit with `.venv/bin/forest repo inspect RUN_ID`. 4. Review execution diagnostics with `.venv/bin/forest run diagnostics RUN_ID`. ### Prompt to paste into your agent ```text Using the FOREST CLI in , clone the public repository into the selected project and wait for it to finish. Then run the inspect and diagnostics commands on the returned run ID and report the source commit. Verify that the commit reported by inspect matches the commit you expected from the repository. ``` ### Caveat Private repositories need worker-side HTTPS or SSH credential profiles, and the README does not describe how those profiles are created, so check the repository authentication guide first. Risk, remote script: Repository code you clone may be executed by the worker under the worker account's permissions, so review the repository before running experiments from it. 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/mikamikasuki-forest/run-experiments-from-public-github-repository/ ## How to verify the exact CSV bytes behind a paired statistics run with Forest Repo: [mikamikasuki/forest](https://github.com/mikamikasuki/forest) | Difficulty: Advanced | Time: 1 hour | Basis: From the README Retrieve the retained input file and provenance record for a paired statistics analysis in FOREST. **Who it is for:** A reviewer or lab member who must confirm which exact data file produced a reported paired statistic. **The problem:** A source CSV may be changed later, so the file on disk no longer proves which bytes were analyzed. **The outcome:** The retained input file, its SHA-256 and size from the run metrics, and a confirmed link between the analysis and the input. ### Steps 1. Submit paired statistics input to `POST /api/projects/{id}/statistics/paired`. 2. Read the run and its input_provenance object with `GET /api/runs/{run_id}`. 3. Download the retained bytes with `GET /api/projects/{id}/download?path=runs/{run_id}/paired-input.csv`. 4. Compare the SHA-256 of the downloaded file with the value in input_provenance. ### Prompt to paste into your agent ```text Using the FOREST API on , fetch run for project , download its retained paired-input.csv, and compute the SHA-256 of the downloaded bytes. Report whether it matches the input_provenance hash in the run record. Verify that the downloaded path is runs//paired-input.csv. ``` ### Caveat The provenance records the bytes the worker consumed, not the bytes at enqueue time, and it is not a general file-version history. 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/mikamikasuki-forest/audit-paired-statistics-input-bytes/