[50 MRG] Speaking path: score from ASR transcript JSON

[TUPM96]

Bounty: 50 MRG

Accept skill=speaking inputs that are transcripts (with optional confidence), produce speaking score + feedback bullets.

Acceptance

• CLI + fixture
• Docs for app integration

Claim (MergeOS MRG)

1. Star https://github.com/mergeos-bounties/NokaMan and https://github.com/mergeos-bounties/mergeos
2. Comment on this issue: I claim this bounty
3. Comment on MergeOS Claim Token #1 with a link to this issue
4. Open a PR to NokaMan (public product repo) with Fixes #<this-issue>

Policy: docs/BOUNTY.md

Important

Work lands on https://github.com/mergeos-bounties/NokaMan — the public product repository.

Payout

Maintainer reviews PR → merge on NokaMan → MRG credit on MergeOS ledger to github:<author> (25/50/100/200 scale).

[Zubi-fix]

/claim #6

Interested in tackling this. I'm reading the codebase and checking for any related tests or edge cases before I start writing. Will drop a PR shortly.

[code-cannot-be-blank]

/attempt #6

I can take this on. Going through the repo structure now and mapping out what needs to change. I'll push a PR when I'm confident the fix is clean and doesn't break anything else.

[samuelpatak-png]

I claim this bounty

[laurentketterle-hub]

I claim this bounty

[laurentketterle-hub]

I claim this bounty (#6)

[laurentketterle-hub]

/claim

[laurentketterle-hub]

/claim