agentleFS
stacklokstacklok/mecatl/.claude/skills/mecak8s-kind-manual-verify/SKILL.md

mecak8s-kind-manual-verify

Skill308 starsChanged 41 days agoApache-2.0View on GitHub

Refresh the mecak8s image in the local mecatl-dev Kind cluster and build mecatui locally, so the user can manually verify in-flight mecatl changes against a real k8s deployment.

  • Reads credentials

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 0 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use this when the user wants to manually exercise in-progress mecatl changes (e.g.

    line 8
  2. Very specific:

    If it doesn't, run task mecak8s:kind-setup first — that's provisioning, not this skill's job.

    line 23
  3. Very specific:

    Use KIND_EXPERIMENTAL_PROVIDER=podman task mecak8s:image-refresh on a Podman host (must match whatever kind-setup was run with).

    line 40
  4. Very specific:

    If the fixture was set up with the optional Keycloak/OIDC identity layer (task mecak8s:kind-keycloak-setup), log in and connect with the …

    line 65

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

mecak8s Kind manual verification loop

Use this when the user wants to manually exercise in-progress mecatl changes (e.g. an open PR branch) against a real mecak8s deployment, rather than only relying on offline tests. This is the "deploy to kind and build mecatui so I can click around" loop — fast iteration, not a fresh cluster bring-up.

This assumes a Kind cluster named mecatl-dev already exists (see deploy/mecak8s-kind/README.md for first-time setup via task mecak8s:kind-setup). This skill is for the iterate step, not initial provisioning.

Prerequisites

  • Working directory is the mecatl repo (or a worktree of it).
  • ko, kubectl, kind, and either docker or podman installed.
  • The mecatl-dev Kind cluster already exists (kind get clusters shows it). If it doesn't, run task mecak8s:kind-setup first — that's provisioning, not this skill's job.

Step 1 — Refresh the mecak8s image with current source

From the repo root:

task mecak8s:image-refresh

This rebuilds ko.local/mecak8s:dev from the current source tree via ko build --local --bare --tags=dev ./cmd/mecak8s, loads it straight into the running mecatl-dev Kind node (no cluster teardown), and does a rolling restart of deployment/mecak8s-mecak8s in the mecatl namespace, waiting up to 240s for the rollout to complete.

Use KIND_EXPERIMENTAL_PROVIDER=podman task mecak8s:image-refresh on a Podman host (must match whatever kind-setup was run with).

This is safe to re-run after every code change you want to test — it does not reset cluster/Redis state.

Step 2 — Build mecatui locally

task build

Produces bin/mecatui (along with bin/mecated, bin/mecademo) via the Taskfile — never a bare go build at the repo root, which drops stray binaries.

Step 3 — Connect and drive it manually

Base fixture (no Keycloak) exposes gRPC at 127.0.0.1:18080 and HTTP at 127.0.0.1:18081 via static Kind NodePort mappings — check with:

task mecak8s:kind-status

If the fixture was set up with the optional Keycloak/OIDC identity layer (task mecak8s:kind-keycloak-setup), log in and connect with the prebuilt shorthand tasks:

task mecak8s:kind-login     # adds /etc/hosts aliases (sudo, once), refreshes fixture CA, runs `mecatui login`
task mecak8s:kind-connect   # adds /etc/hosts aliases, refreshes fixture CA, runs `mecatui connect`

Both require ./bin/mecatui to already exist (Step 2) and task mecak8s:kind-keycloak-setup to have been run at least once. See the keycloak-kind-setup skill if the Keycloak layer itself needs (re-)provisioning or debugging.

Without the Keycloak layer, connect directly:

./bin/mecatui connect 127.0.0.1:18080

For identity broker or hosted AS-1 work, use the separate mecatl-idbroker profile instead (Dex upstream, mecatui login through AS-1; runs alongside mecatl-dev): task mecak8s-idbroker:kind-setup, then task mecak8s-idbroker:image-refresh to iterate. See deploy/mecak8s-idbroker/README.md.

Iterating

Repeat Step 1 (task mecak8s:image-refresh) after each code change, then reconnect with mecatui (Step 3) — no need to rebuild mecatui itself unless the TUI/client code changed.

Cleanup

This skill never tears the cluster down. To destroy the whole fixture:

task mecak8s:kind-destroy

Reference

  • deploy/mecak8s-kind/Taskfile.yml — image-refresh, kind-status, kind-login, kind-connect, kind-destroy task definitions.
  • deploy/mecak8s-kind/README.md — full fixture lifecycle, provider modes (mock vs. real via OPENROUTER_API_KEY), and the optional Keycloak layer.
  • keycloak-kind-setup skill — for provisioning/recovering the Keycloak layer itself, if this fixture uses it.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.