# AI use cases for Agent Foundation (converge-ai-labs/agent-foundation) Repo: [converge-ai-labs/agent-foundation](https://github.com/converge-ai-labs/agent-foundation) | Use cases: 4 | Page: https://dedupedai.com/use-cases/converge-ai-labs-agent-foundation/ Agent Foundation (a13n) is an open-source, self-hosted library and platform for building and running your own AI agent systems, with managed agents, memory, sandboxes, computer use, and durable execution. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Run your own agent service and try an agent: https://dedupedai.com/use-cases/converge-ai-labs-agent-foundation/self-host-agent-service-with-compose/ - Embed reusable agents in your own application: https://dedupedai.com/use-cases/converge-ai-labs-agent-foundation/embed-harness-runtime-in-your-app/ - Try an agent on your own project in a terminal: https://dedupedai.com/use-cases/converge-ai-labs-agent-foundation/try-agents-on-your-repo-in-terminal/ - Review agent failures and approve proposed fixes: https://dedupedai.com/use-cases/converge-ai-labs-agent-foundation/turn-agent-failures-into-reviewed-fixes/ ## How to run your own agent service and try an agent with Agent Foundation Repo: [converge-ai-labs/agent-foundation](https://github.com/converge-ai-labs/agent-foundation) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Start a self-hosted agent service with Docker Compose and send a first message. **Who it is for:** A developer or platform owner who wants agents running on their own servers instead of a hosted vendor, starting from a machine with Docker installed. **The problem:** Building an agent backend from scratch means wiring models, execution environments, and state before any agent can run. **The outcome:** A running Service and Console with an administrator account, a connected model, and an agent that answers a first message. ### Steps 1. `mkdir a13n-service` 2. `cd a13n-service` 3. `curl -fL https://raw.githubusercontent.com/converge-ai-labs/agent-foundation/main/deploy/docker/compose/a13n-service.yaml -o a13n-service.yaml` 4. `docker compose -f a13n-service.yaml up -d --wait --pull always` 5. Open the Console, register the administrator account with a password of at least 8 characters, then add a model under Models using an API key or Sign in with ChatGPT. 6. Create an agent, choose the model, and send a message with Try agent. ### Prompt to paste into your agent ```text In the agent-foundation repo (converge-ai-labs/agent-foundation), follow the Docker Compose steps in the README to start the Service stack in an empty . Report the Console URL and confirm that the containers reported healthy by `docker compose` came up without errors. ``` ### Caveat The stack binds to loopback and mounts the host Docker socket, so the README says a shared deployment needs its separate HTTPS and bootstrap steps. Risk, credentials: Model provider API keys or ChatGPT sign-in are entered into the Console, so store them only on a machine you control. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. 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/converge-ai-labs-agent-foundation/self-host-agent-service-with-compose/ ## How to embed reusable agents in your own application with Agent Foundation Repo: [converge-ai-labs/agent-foundation](https://github.com/converge-ai-labs/agent-foundation) | Difficulty: Advanced | Time: 60 minutes | Basis: From the README Install the Harness library from source and start its quickstart. **Who it is for:** An application developer who wants agents with typed tools, streaming output, and resumable state inside their own product. **The problem:** Agent logic tends to get tangled with product code, which makes tools, outputs, and saved state hard to reuse. **The outcome:** A local checkout with the Harness package synced, ready for the quickstart and runnable examples. ### Steps 1. `git clone https://github.com/converge-ai-labs/agent-foundation.git` 2. `cd agent-foundation` 3. `uv sync --locked --package a13n-harness` 4. Follow the Harness quickstart at docs/a13n-harness/getting-started.md. 5. Open examples/README.md and run one example that matches your application. ### Prompt to paste into your agent ```text In the agent-foundation repo, clone it, run `uv sync --locked --package a13n-harness`, and report the sync result. Then list the runnable examples in examples/README.md that use the Harness package and name one that fits . ``` ### Caveat The README says your application owns persistence, credentials, and recovery policy, so those parts are not provided by the library. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. 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/converge-ai-labs-agent-foundation/embed-harness-runtime-in-your-app/ ## How to try an agent on your own project in a terminal with Agent Foundation Repo: [converge-ai-labs/agent-foundation](https://github.com/converge-ai-labs/agent-foundation) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Install Harness UI and run its terminal agent inside a repository. **Who it is for:** A developer who wants to test models, tools, and Skills against their current project without building an application first. **The problem:** Comparing models and agent configurations usually means writing glue code before any real output appears. **The outcome:** A terminal agent session in your repository, with a connected model and chosen execution permissions. ### Steps 1. `uv tool install a13n-harness-ui` 2. `cd your-repository` 3. `a13n-harness-ui` 4. Complete first-use setup to connect a model and select execution permissions. 5. Ask the agent to explore a file in the repository and review its output. ### Prompt to paste into your agent ```text Install a13n-harness-ui with `uv tool install a13n-harness-ui`, then run `a13n-harness-ui` inside . Report the model you connected and the execution permission you chose, and confirm the agent can read one file you name. ``` ### Caveat Full Control runs as your host account, so pick a narrower execution permission for unfamiliar repositories. Risk, credentials: First-use setup asks for a model connection, so use a key scoped for testing and avoid pasting it into shared sessions. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. 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/converge-ai-labs-agent-foundation/try-agents-on-your-repo-in-terminal/ ## How to review agent failures and approve proposed fixes with Agent Foundation Repo: [converge-ai-labs/agent-foundation](https://github.com/converge-ai-labs/agent-foundation) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Inspect execution evidence in Findings and ask Composer for a reviewed change. **Who it is for:** An operator running agents in Service who needs to find why runs failed and decide what to change. **The problem:** Failed agent runs are hard to diagnose, and fixes applied without review can make things worse. **The outcome:** A recorded diagnosis and a proposed Agent revision that you reviewed and approved before anything was published. ### Steps 1. Start the stack with `docker compose -f a13n-service.yaml up -d --wait --pull always`. 2. Open Findings in the Console and review the scanned diagnoses and their status. 3. Inspect the evidence for one diagnosis and record your judgment. 4. Send the cited Agent revision and evidence to Composer to request a proposed change. 5. Review the proposed change and approve or reject it. Nothing is repaired or published automatically. ### Prompt to paste into your agent ```text In the agent-foundation Service Console, open Findings and list the diagnoses with status for . Pick one, summarize its cited evidence, and confirm that no change was published before you review it. ``` ### Caveat Findings depends on existing execution evidence, so it only helps after agents have run through a Service stack that is recording traces. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. 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/converge-ai-labs-agent-foundation/turn-agent-failures-into-reviewed-fixes/