Skip to main content

Generic Agent + Summarizer Skill Demo in the Rossoctl UI

This guide explains how to use the Rossoctl UI to:

  1. import the example summarizer skill,
  2. import the example generic_agent agent,
  3. link the skill to the agent from the UI,
  4. and verify in chat that the skill is available and being used.

This flow is intended for the Rossoctl UI workflow where skills are imported into Rossoctl first and then linked to an agent during agent import.

What this demo shows​

The example generic agent loads skill instructions and exposes them through its agent card in the Rossoctl UI. In this demo:

  • the skill content comes from the imported summarizer skill,
  • the agent is the example generic_agent,
  • the linkage between the two is configured in the Rossoctl UI when the agent is imported,
  • and the result is visible both in the agent card and in the agent chat experience.

Prerequisites​

Before starting, make sure:

  • Rossoctl is installed and the UI is reachable, as described in docs/install.md

  • the Skills feature flag is enabled — Skills are disabled by default and must be explicitly enabled during installation (see Enabling Skills)

  • the build system is deployed — Agent builds require Shipwright and the in-cluster registry (registry.cr-system.svc.cluster.local:5000). Deploy with --with-builds or the build push step will fail with no such host. If you need to add it to an existing cluster: scripts/kind/setup-rossoctl.sh --with-builds --skip-cluster

  • you have access to a Rossoctl-enabled namespace, for example team1

  • the cluster can build example agents from GitHub

  • you have LLM credentials ready for the generic agent

  • featureFlags.skills is enabled — use --with-skills with the Kind setup script:

    scripts/kind/setup-rossoctl.sh --with-skills --with-builds

    Or enable on a running cluster:

    helm upgrade rossoctl ./charts/rossoctl/ \
      --reuse-values \
      --set featureFlags.skills=true \
      -n rossoctl-system

For the example generic agent, configure the LLM environment variables with values appropriate for your setup. At minimum, the generic agent needs:

  • LLM_MODEL
  • LLM_API_BASE
  • LLM_API_KEY

If you use the example env file flow in the UI, you can import the matching env file from the example repository and then adjust values if needed.

Repositories and paths used in this demo​

This guide uses the example repository:

  • Repository URL: https://github.com/rossoctl/examples
  • Skill path: skills/summarizer
  • Agent path: a2a/generic_agent

Step 1: Import the summarizer skill in the UI​

  1. Open the Rossoctl UI.

  2. Navigate to Skills.

  3. Click Import Skill.

  4. In Namespace, select the namespace you will also use for the agent, for example team1.

  5. In Name, enter summarizer.

  6. In Description, enter a clear description, for example:

    Summarization skill for converting long source text into concise structured summaries.

  7. In URL, paste the example skill URL:

    https://github.com/rossoctl/examples/tree/main/skills/summarizer

  8. Wait for the UI to auto-import the skill content and additional files.

  9. Confirm that SKILL.md was imported.

  10. Click Import Skill.

After the skill is created, the skill should appear in the skill catalog for the selected namespace.

Step 2: Import the generic agent in the UI​

  1. Navigate to Agents.

  2. Click Import New Agent.

  3. In Namespace, select the same namespace where you imported the skill.

  4. Leave Deployment Method as Build from Source.

  5. In Git Repository URL, use:

    https://github.com/rossoctl/examples

  6. In Git Branch or Tag, use main.

  7. In Select Agent, choose Generic Agent. This fills the source path with:

    a2a/generic_agent

  8. Confirm the generated or entered agent name. For this demo, generic-agent is a good choice.

  9. Set Protocol to a2a.

  10. Set Framework to LangGraph.

Step 3: Configure the generic agent environment variables​

In the Environment Variables section, provide the LLM configuration required by the generic agent.

Add these variables with values that match your environment:

  • LLM_MODEL
  • LLM_API_BASE
  • LLM_API_KEY

Notes:

  • If you are using a hosted model endpoint, set these values according to that provider.
  • If you are using a local model setup supported by Rossoctl, use the values documented in docs/local-models.md.

Do not set SKILL_FOLDERS manually in this demo. When you select a skill in Linked Skills, Rossoctl mounts the linked skill ConfigMaps into the agent pod under /app/skills/... and sets SKILL_FOLDERS automatically for the generic agent runtime.

In the Linked Skills section of the import form:

  1. Find the imported summarizer skill in the list of namespace skills.
  2. Enable the checkbox for summarizer.

This stores the skill linkage with the agent configuration so Rossoctl can associate the imported skill with the imported generic agent, mount the skill files into the pod, and expose those mounted paths to the generic agent through SKILL_FOLDERS.

Step 5: Build and deploy the agent​

  1. Review the remaining defaults.
  2. Click Create / Build & Deploy for the agent.
  3. Wait for the Shipwright build to complete.
  4. Open the agent details page when the deployment finishes.

Step 6: Verify that the skill is linked​

On the agent details page:

  1. Confirm the agent status is healthy.
  2. Locate the agent card or skills section.
  3. Verify that the summarizer skill appears in the agent’s listed skills.
  4. If you inspect the running pod, verify Rossoctl populated SKILL_FOLDERS automatically and mounted the skill under /app/skills/summarizer.

This confirms both the UI-visible skill linkage and the runtime skill wiring between the imported skill and the imported generic agent.

Step 7: Open the chat and test the summarizer skill​

From the agent details page, open the chat UI for the generic agent.

Use a prompt that strongly signals summarization. Paste the following demo text into the chat:

Use your summarizer skill to summarize the following project update into:
1. a one-sentence executive summary,
2. exactly 5 bullet points,
3. a short risk list,
4. and 3 clear action items.

Project update:
During the last two sprints, the platform team completed the first end-to-end integration between the Rossoctl UI and the example generic agent. The team also imported the summarizer skill into the namespace and linked it to the agent during the UI import flow. Initial testing showed that the agent can accept long-form text and respond with a concise structured summary. However, several follow-up items remain: the team needs to improve documentation, verify the build flow in a fresh namespace, and confirm that the agent card correctly displays linked skills after deployment. There is also an open concern that users may forget to provide the required LLM environment variables, which leads to startup failures that are not always obvious from the UI alone. If the remaining validation passes, the team plans to use this demo in the next stakeholder walkthrough to show how Rossoctl can manage both reusable skills and example agents through the same UI.

Expected result​

A successful response should look like a structured summary, not a free-form essay.

You should expect the answer to include:

  • one short executive-summary sentence,
  • 5 bullets summarizing the update,
  • a risk section,
  • and 3 action items.

A representative successful output would look like this:

Executive summary:
The team successfully connected the Rossoctl UI, the generic agent, and the summarizer skill, and now needs to complete validation and documentation before using the flow in a stakeholder demo.

Key points:
- The team completed an end-to-end integration between the Rossoctl UI and the generic agent.
- The summarizer skill was imported into the namespace and linked during agent import.
- Initial testing showed the agent can summarize long-form text into a concise structure.
- Documentation and fresh-namespace validation are still pending.
- Missing LLM environment variables remain a usability risk during startup.

Risks:
- Users may omit required LLM configuration.
- Fresh-environment validation may reveal deployment issues.
- Skill display in the agent card still needs confirmation.

Action items:
1. Finalize the step-by-step documentation.
2. Validate the full flow in a new namespace.
3. Confirm agent-card skill visibility before the stakeholder demo.

The wording does not need to match exactly, but the structure and behavior should clearly reflect summarization rather than general chat.

How to tell that the skill is working​

The skill is working if all of the following are true:

  • the summarizer skill is visible in the agent card or skill list,
  • the agent responds to the long-form prompt with a concise structured summary,
  • the response follows the requested summary structure,
  • the runtime did not require you to manually set SKILL_FOLDERS,
  • and the response is clearly based on summarization behavior rather than a generic unconstrained answer.

For a live demo, use this sequence:

  1. Show the Import Skill page and import the summarizer skill from GitHub.
  2. Show the Import New Agent page and select a2a/generic_agent.
  3. Show the Linked Skills section and select summarizer.
  4. Show the finished agent detail page where the skill is listed.
  5. Open chat and paste the long project-update prompt.
  6. Point out that the resulting response is structured as a summary, which demonstrates the linked skill flow.

Troubleshooting​

The Skills section is not visible in the UI​

The Skills feature is disabled by default. If the Skills navigation item or the Import Skill button is missing, the feature flag was not enabled at install time. See Enabling Skills for how to enable it with the setup script (--with-skills) or how to enable it on an existing cluster with helm upgrade without a full redeploy.

The build fails with no such host when pushing the image​

The in-cluster registry was not deployed. Re-run setup with --with-builds:

scripts/kind/setup-rossoctl.sh --with-backend --with-ui --with-skills --with-builds --build-images --skip-cluster

--with-builds deploys Shipwright and the container registry at registry.cr-system.svc.cluster.local:5000 that agent builds push to.

The Linked Skills section is missing from the Import Agent form​

The Linked Skills section was added after the v0.7.0-alpha.2 release. If the running UI image predates that, the section will not appear at all. Rebuild the UI from local source and reload it into Kind:

export ROSSOCTL_FEATURE_FLAG_SKILLS=true
scripts/kind/setup-rossoctl.sh --with-backend --with-ui --with-skills --build-images --skip-cluster

"Agent already exists" error when reimporting after a failed build​

If a previous import attempt failed mid-way (for example, the build pod errored out), leftover Shipwright and Kubernetes resources remain in the namespace and block a fresh import. Delete them before retrying, replacing <agent-name> and <namespace> with your values:

kubectl delete buildrun -n <namespace> -l app=<agent-name>
kubectl delete build <agent-name> -n <namespace>
kubectl delete configmap <agent-name>-card-unsigned -n <namespace>

Then retry the import in the UI.

The summarizer skill does not appear in the Linked Skills section​

Check that:

  • the skill was imported successfully,
  • the skill was imported into the same namespace as the agent,
  • and the skill appears in the skill catalog before opening the agent import page.

The agent deploys but chat responses are poor or fail​

Check that:

  • LLM_MODEL, LLM_API_BASE, and LLM_API_KEY are set correctly,
  • the model endpoint is reachable from the agent,
  • and the agent pod started successfully.

The agent does not show the linked skill on its detail page​

Check that:

  • the skill was selected in Linked Skills during agent import,
  • the agent build and deployment completed successfully,
  • and the agent card was generated for the deployed agent.

The skill is listed but still does not seem to load at runtime​

Check that:

  • the linked skill name matches the imported skill name in the same namespace,
  • the agent pod has SKILL_FOLDERS set,
  • the agent pod has mounted skill content under /app/skills,
  • and the imported skill includes the expected files such as SKILL.md.

The answer is not obviously summarized​

Use a longer prompt and explicitly ask the agent to use the summarizer skill and return a structured format. The provided demo prompt is designed for that purpose.

Cleanup​

After the demo, delete the agent and skill from the Rossoctl UI:

  1. Go to Agents and delete the generic agent.
  2. Go to Skills and delete the summarizer skill.