BuilderIO/agent-native / .agents/skills/verifying-changes
verifying-changes
VerifiedConcrete, per-area proof that a change actually works before reporting it fixed, done, or "should work now" — which dev server, test command, or invocation proves a template UI change, an action, a migration, a guard, or a core/package change. Use before every wrap-up, and before stopping mid-task to ask permission instead of continuing.
Verifying Changes
When this applies
Before telling the user a fix, feature, or bug is done, choose the smallest proof that exercises the behavior you changed and use the row below as a guide. "I changed the code and it looks right" is not proof — it's the exact gap this skill exists to close. A failing check is a reason to keep working, not a reason to stop, ask, or report done anyway.
Scale the verification
Verification proves the changed behavior; it is not a reason to run a broad workspace audit after every edit.
- For a localized one-line or small-file change, inspect the diff and run the
narrowest relevant test, typecheck, formatter, or direct invocation. Do not
run
pnpm run prep, browser automation, or restart a dev server unless the changed area requires it. - For an action or capability-card change, use the focused action/card check. Restart only when the runtime process must reload registration or the change explicitly affects startup.
- Escalate to package-wide tests or
pnpm run prepfor shared contracts, cross-cutting changes, migrations, or when a focused check exposes a wider failure. Keep repository-required guards and doctor checks when they apply. - If a focused check is unavailable, state the exact gap rather than replacing it with unrelated expensive work.
Do NOT report done in these situations
- The change is "obviously correct" or one line — obvious fixes are the ones that ship broken most often; run the smallest relevant proof anyway.
- You verified a similar path earlier in the session — re-run against the actual latest edit, not a memory of an earlier pass.
- The user didn't explicitly ask you to test it — test it anyway; verify- before-done is a standing rule here, not an opt-in.
- The full test suite is slow or flaky — run the targeted command for the changed area (below) instead of skipping verification entirely.
- You're mid-task and unsure whether to keep going — keep going. Only stop for a missing credential, an ambiguous decision only the user can make, or a destructive action that needs confirmation. Silence from the user means keep working, not pause and wait.
- You genuinely cannot run anything — say so out loud (see below); never let "unverified" read as "done".
Proof by area
| You changed | What proves it | Command |
|---|---|---|
Template UI/behavior (templates/<app>/app/**) | Drive the real page and check console + network, not just the diff | pnpm --filter <app> dev (or root pnpm dev for the gateway), then click through the exact flow with whatever browser tool is available; check for console errors and failed (4xx/5xx) requests on that page. A sign-in screen is not a blocker: POST /_agent-native/auth/local-dev from the page's own context (Playwright page.request.post) signs in as the local dev account |
An action (templates/<app>/actions/*.ts) | Call it with representative args and inspect the real return value | cd templates/<app> && pnpm action <name> --key value; for a write, follow with pnpm action db-query --sql "SELECT ..." to confirm the row actually landed |
| Schema/migration | Boot the app so migrations run, then read back the new column/table | pnpm --filter <app> dev once, then cd templates/<app> && pnpm action db-query --sql "..." (action is a per-template script, not a root one); pnpm guard:additive-migrations catches destructive DDL before CI does. For a core store: pnpm gen:store-registry, then pnpm guard:release-schema-complete |
A guard/lint script (scripts/guard-*.{mjs,ts}) | Run it directly against a case that should now pass and one that should still fail | pnpm guard:<name> (name matches the package.json script); pnpm guards for the full sweep |
packages/core or another publishable package | Run that package's actual tests, not just typecheck | pnpm --filter @agent-native/core exec vitest --run <changed.spec.ts>, or pnpm test:core-integration for cross-cutting paths |
A CI job, workflow, or scripts/ci-*.ts | Show the new scope on real history; the PR's own CI is a full run and cannot show it | node --experimental-strip-types --test scripts/ci-change-scope.test.ts, then run scripts/ci-change-scope.ts with GITHUB_OUTPUT on real main commits; see adding-tests-and-ci |
A deploy guide (node-js.mdx, docker.mdx, netlify.mdx, vercel.mdx) or a package a standalone Chat app installs | Run the documented path: scaffold, migrate a real Postgres, build, run, sign up, restart | pnpm qa:deploy-paths --target node, docker (Docker required; it starts its own Postgres), or presets (the Netlify and Vercel build output). --target netlify or vercel deploys to a real an-deploy-ci-<host>-<slot> project, created once and reused, and needs DEPLOY_CHECK_DATABASE_URL plus that host's login (--netlify-account-slug, --vercel-scope). --source npm proves what is published instead of this checkout |
How fast a new workspace gets created (packages/core/src/cli/create.ts, git-http.ts, or a create package) | Compare cold runs against what npm users get now: time, bytes, and connections per phase | pnpm bench:create compares npm latest with this checkout, packed. Add --target <spec or .tgz> to compare others, --runs 3 before quoting times, and --pnpm-install to include the workspace's install. --standalone measures the Getting Started path; --add-app <template> also times adding an app to the new workspace |
| Cross-cutting change, or unsure which area | Workspace-wide pass | pnpm run prep (fmt + typecheck + test:fast + guards, run in parallel) |
| Deployed behavior (when requested or environment-specific) | Add a deployed check only when requested or a concrete beta/production difference could make local proof misleading (for example, deploy-time config or secrets, host/cookie/OAuth settings, CDN/SSR, serverless runtime, storage, or external connectivity). A merge-triggered deploy alone is not a reason to retest every path remotely; focused local proof and required CI are the default. When warranted, exercise the affected path on that environment and inspect relevant logs; deploy success alone is not behavior proof | |
Docs only (.md, AGENTS.md, SKILL.md) | Nothing to run | Say "docs-only, no runtime check applies" — don't invent a verification step |
For any user-visible change, put the proof in the reply: a screenshot of the surface you just drove, or the actual query result / log line for backend work. "Show me screenshots" is a standing expectation, not a special request.
pnpm test:fast excludes .db.test.ts / .integration.* / .e2e.* /
.live.* / .perf.* suites. If your change touches one of those, name and
run that specific file — test:fast passing does not cover it.
Production forensics
When inspecting production runs, query interactive and scheduled/background
work as separate populations before summarizing reliability. Report both
id NOT LIKE 'job-%' and id LIKE 'job-%' (or the repo's current equivalent),
including app, run count, completed count, failure count, and top terminal
reasons for each slice. A healthy interactive sample does not prove scheduled
jobs work.
When you can't verify
State it plainly and name what would close the gap: "I could not run this —
verifying it needs <command> or a browser check of <page>." Never write
"should be fixed" or "this resolves it" without having actually run the check
above.
Real failures this replaces
- "still getting it friend. this is the third time you said you fixed it when you didn't. please reproduce end to end and verify"
- "did you test end to end? can you do so in the browser and confirm?" — asked on nearly every wrap-up before this rule existed
- "ok so should work now?? i am getting sick of saying 'try this' and it still not working. you confident?"
- "my analytics dashboards ALWAYS fail ... i have asked agents to fix this for weeks at least 10x and they always say they did and then the emails keep failing"
- "WHY THE FUCK DO YOU KEEP STOPPING" / "sorry what is still queued? you should be doing everything now don't queue" — stopping mid-task instead of finishing and verifying
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the work | Stars: 319k | Updated | Claude CodeCodex | Verified | Install |
| systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixes | Stars: 319k | Updated | Claude CodeCodex | Verified | Install |
| test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation code | Stars: 319k | Updated | Claude CodeCodex | Verified | Install |
| skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |