Skip to content

ONBOARD_TESTSPRITE.md

Latest commit

 

History

History
108 lines (76 loc) · 3.7 KB

ONBOARD_TESTSPRITE.md

File metadata and controls

108 lines (76 loc) · 3.7 KB
Raw

Onboarding TestSprite — Mestre do PC V10

This repo currently has no TestSprite tests. The artefacts below are ready to be uploaded once the TestSprite CLI is available and the launcher is running.

1. Start the launcher

Open PowerShell as Administrator in Mestre-do-PC-V10-repo-pronto/Mestre-do-PC-V10-clean/:

.\MestreDoPC-Launcher.ps1

Wait for:

[OK] Aguardando comandos do HTML...

2. Ensure testsprite setup has run

testsprite project list

If this errors on auth, run testsprite setup first.

3. Create projects

Frontend project

cd Mestre-do-PC-V10-repo-pronto/Mestre-do-PC-V10-clean
testsprite project create --type frontend --name "Mestre do PC V10" --url http://127.0.0.1:7777/

Capture the returned projectId (e.g. proj_abc123) and replace __FE_PROJECT_ID__ in every file under testsprite-plans/*.json.

Backend project

testsprite project create --type backend --name "Mestre do PC V10 Backend"

Capture the returned projectId (e.g. proj_def456).

Configure the static credential so TestSprite injects __AUTH_HEADERS__:

testsprite project credential <BE_PROJECT_ID> --type "API key" --credential "v10-web"

Then update the backend scripts (test_run_safe_command.py, test_run_blocked_command.py, test_open_terminal_rejects_unauthorized.py) to spread __AUTH_HEADERS__ instead of the local AUTH_HEADERS dict. Example:

r = requests.post(
    f"{TARGET_URL}/run",
    headers={**__AUTH_HEADERS__, "Content-Type": "application/json"},
    json=payload,
    timeout=10,
)

4. Create tests

Frontend (batch):

testsprite test create-batch --plan-from-dir ./testsprite-plans

Backend (one per file):

testsprite test create --type backend --name "ping returns ok" --code-file ./testsprite-backend/test_ping.py --project <BE_PROJECT_ID>
testsprite test create --type backend --name "status returns system metrics" --code-file ./testsprite-backend/test_status.py --project <BE_PROJECT_ID>
testsprite test create --type backend --name "mcp status is reachable" --code-file ./testsprite-backend/test_mcp_status.py --project <BE_PROJECT_ID>
testsprite test create --type backend --name "ollama tags returns model list" --code-file ./testsprite-backend/test_ollama_tags.py --project <BE_PROJECT_ID>
testsprite test create --type backend --name "run safe readonly command" --code-file ./testsprite-backend/test_run_safe_command.py --project <BE_PROJECT_ID>
testsprite test create --type backend --name "run rejects unallowed command" --code-file ./testsprite-backend/test_run_blocked_command.py --project <BE_PROJECT_ID>
testsprite test create --type backend --name "open terminal requires auth header" --code-file ./testsprite-backend/test_open_terminal_rejects_unauthorized.py --project <BE_PROJECT_ID>
testsprite test create --type backend --name "static assets served with cache headers" --code-file ./testsprite-backend/test_static_assets.py --project <BE_PROJECT_ID>

5. Smoke-run a few (protect credits)

Pick 2–3 happy-path tests, for example:

testsprite test run <testId-ping> --wait
testsprite test run <testId-status> --wait
testsprite test run <testId-fe-server-status> --wait

Do not run the full 16-test suite automatically — it would cost ~30–35 credits. Running the rest is the user's choice.

Test summary

  • Frontend: 8 plans covering dashboard status, command execution, search, favorites, IA chat, copy-to-clipboard, sidebar navigation and theme toggle.
  • Backend: 8 Python tests covering /ping, /status, /mcp-status, /ollama/tags, /run (happy + rejection), /open-terminal auth enforcement and static-asset cache headers.

Total: 16 tests.