[Pavithran-R-A]

Goal

Turn MergeSutra's 2,192-line, implementation-diary-style README into a crisp, evidence-backed landing page for BharatCode Build League judges, developers and maintainers.

Result

• 167-line product page (down from 2,192) with the hook AI can propose a fix. MergeSutra demands evidence.
• Demonstration and a judge's guide accessible immediately.
• Problem → solution → what makes it different → evidence → install → safety boundaries.
• Honest disclosure that the demo uses a scripted gateway, the older live-Qwen run is historical, current-head live E2E and ten-case benchmark are still missing, and IDE integration is not shipped.
• Direct calls to action for authentic feedback, votes and stars; no artificial engagement.
• All model/configuration variables, OS matrix claims and package-safe README pointers preserved.

Tests / review constraints

• Keep tests/security/promise-vocabulary.test.ts green (README opening promise measured at 645 characters; allowed 800).
• Preserve tests/config/config-docs.test.ts, tests/docs/platform-matrix.test.ts, tests/security/shipped-pointer-boundary.test.ts, package and artifact tests.
• Require all 6 CI matrix jobs + 8 scripted customer journeys.
• No code, runtime, security policy, API key, model spending or remote publication modifications.

Relevant baseline

main at c6fd87102d24977b255bc4bbb7fb2a3e1c2cf24e, with 6/6 green CI: https://github.com/Pavithran-R-A/mergesutra/actions/runs/42492506208.

This PR does not itself prove the present SHA's live BharatCode issue-to-PR execution.

[Copilot]

Copilot wasn't able to review any files in this pull request. Copilot's per-file or total diff limits were exceeded. Try reducing the changed content or splitting the changes into smaller pull requests, then request a review from Copilot again.

[linear-code]

PAV-162