ai style

Author: LouiseMedova

script/analyze-testnet-ts-run.mjs

@@ -55,15 +55,17 @@ You are analyzing a GitHub Actions run for this repository.
 
 Return a concise plain-text report in exactly this format:
 Summary: <one short paragraph>
-Probable root cause: <one short paragraph>
-Suggested next step: <one short paragraph>
+Risk: <one short paragraph>
+Next: <one short paragraph>
 
 Rules:
 - Be concrete and repository-specific.
 - If the run succeeded, say that clearly and mention any residual risk.
 - If the run failed, identify the most likely failing layer: config, CI bootstrap, RPC/network, test logic, or flaky external dependency.
 - Prefer evidence from the log over speculation.
-- Keep the whole report under 140 words.
+- Optimize for Telegram readability.
+- Keep the whole report under 90 words.
+- Each line should be compact and easy to scan on mobile.
 - Do not use Markdown headings, bullets, numbering, or code fences.
 
 Repository rules: