upgrade-canarylisted

Re-verify that Claude Code's Workflow-tool runtime still behaves the way the @workflow-toolbox toolkit depends on, after a Claude Code upgrade — and report what changed. Invoke when the user upgraded (or might have upgraded) Claude Code and asks whether the workflow toolkit / committed artifacts still work, when they say "run the upgrade canary", "did the update break the workflows", "re-check the runtime surface", "what changed in the runtime", or before cutting a plugin release. The canary is version-triggered: the full check runs only when the `claude` CLI or the Agent SDK version changed since it last passed (or when forced), so it is cheap to invoke often. It tests TWO runtimes — the user's interactive `claude` binary AND the SDK-bundled one — and reports their version delta + any behavior drift, which can drive fixes or feature work. This is a MAINTAINER tool — it runs against this repo's `toolkit/` dev tree and needs local Claude Code subscription auth; it is not meant for end users who only installed
home-dev-lab/workflow-toolbox · ★ 0 · AI & Automation · score 63

Install: claude install-skill home-dev-lab/workflow-toolbox

# Upgrade canary — re-verify the runtime after a Claude Code upgrade Claude Code's Workflow tool is a **research preview**: parts of the surface the `@workflow-toolbox` toolkit relies on (the 512 KB script cap, the "`meta` must be the first statement" rule, the bundled-artifact launch path) are locally-verified, not publicly contracted, and can drift on an upgrade. This canary re-checks that surface, on **both runtimes you actually use**, so a breakage — or a behavior change worth adopting — is surfaced here, deliberately. It is **version-triggered, not scheduled**. Claude Code updates almost daily; the canary first asks "did the runtime change since it last passed?" and only does the expensive work when the answer is yes (or you pass `--force`). > **Maintainer tool.** Every command below runs from this repo's `toolkit/` > directory and drives the real runtime via the Agent SDK under your local Claude > Code subscription (the SDK reuses `~/.claude` credentials — no API key). The > canary harness lives in `@workflow-toolbox/smoke`, a **private** workspace > package that is not published to npm — so unlike `workflow-toolbox scaffold/build/check/ > debug/report`, this skill is **not expected to run from a consumer project**. > It is not runnable from an end-user plugin install either, which ships only the > built artifacts, not `toolkit/`. If `toolkit/` is absent, say so and stop. ## The two runtimes `@workflow-toolbox` workflows run in **two** Claude Code binaries that dri