Qwen and Alibaba readiness without live secrets

PrizePilot proves live Qwen refinement and keeps endpoint proof gated.

This page is the public, no-login evidence packet for sponsor proof. It verifies the Qwen request shape, runtime-only key boundary, completed Qwen live smoke proof, prepared Alibaba Function Compute code path, dashboard payload, and public claim boundary without publishing secrets or private account data.

5+2

five completed checks plus two prepared Alibaba proof targets generated by python -m prizepilot cloud-readiness.

0

secrets, cookies, account IDs, billing data, payout data, tax data, or KYC fields required.

1 gate

live Alibaba endpoint remains an account-owner action; Qwen live smoke proof is captured.

Judge Artifacts

Markdown Report

Readable evidence table with pass, source-code-prepared, prepared-not-live, and exact next live evidence steps.

JSON Report

Machine-readable report for reviewers who want the current readiness status directly.

Live Proof Gate

Action-time Qwen/DashScope and Alibaba proof checklist with secret and billing boundaries.

Qwen Live Proof

Non-sensitive runtime smoke proof for the completed Qwen/DashScope refinement path.

Judge Evidence Pack

Track 4 matrix, public proof, screenshots, and account-gated evidence boundary.

Verification Matrix

Check Status What it proves Live step still required
Qwen-compatible request shape Pass src/prizepilot/qwen_client.py uses DashScope compatible mode and chat-completions payload fields. Run the same path with a runtime-only key supplied by the account owner.
Runtime-only secret boundary Pass Keys are read from environment variables and are not written to the public ledger. Remove the environment variable after the live smoke test and publish only non-secret excerpts.
Qwen live smoke proof Pass A runtime-only Qwen/DashScope call using qwen-plus returned a refined Blog Post Award plan and confirmed the key was absent after cleanup. Future reruns still require account-owner key entry at action time.
Alibaba Function Compute manifest Source code prepared deploy/alibaba-cloud/s.yaml defines a custom-container Function Compute service for PrizePilot. Deploy after account-owner approval for region, credit, and billing implications.
Public HTTP proof targets Prepared, not live The manifest exposes an HTTP trigger and health check path suitable for / and /api/plan verification, but no live Alibaba URL is claimed here. Verify the real endpoint URL and record HTTP 200 evidence.
Dashboard judge payload Pass The local payload includes the submitted Qwen target, status, scorecard, approval queue, and evidence gaps. Use the same payload as the expected success signal after Alibaba deployment.
Public claim boundary Pass The project clearly states live Qwen proof is captured and live Alibaba endpoint proof is not claimed yet. Update Devpost only after endpoint proof exists.
Honest sponsor signal: this is not a substitute for live Alibaba endpoint proof. It is the reproducible public certificate that Qwen live proof is captured and the remaining endpoint gate is bounded before the account owner approves deployment.

Regenerate

From the repository root:

$env:PYTHONPATH='src'
python -m prizepilot cloud-readiness
python -m prizepilot cloud-readiness --checked-at 2026-06-15T12:30:00Z
python -m prizepilot qwen-status
python -m unittest discover -s tests -v

These commands do not require a Qwen key, Alibaba login, billing data, or any private account storage. Use --checked-at only for deterministic evidence rebuilds and diffs; live public updates should regenerate the report with the current action-time timestamp.