gm
gm

Code Judgment Endpoint

Cross-model DoD verdict for a PR. Called by Platform EM before merge decision.

Runs:

  1. Mechanical DoD Checker (deterministic, ~50ms, no LLM)
  2. Code Judgment Agent (OpenAI single-shot, ~3-8s, different model family from Platform EM)
  3. Aggregator (combines verdicts into final decision per tier rules)

Returns an AggregateVerdict as JSON. Platform EM interprets:

  • decision=auto_merge_eligible → proceed with gh pr merge
  • decision=advisory_pass → comment only, advisory review
  • decision=requires_human_ack → Tier S; route to Google Doc product-level ack
  • decision=block → do not merge; block_reasons explain why
  • decision=degraded → judgment unavailable; handle per tier

The tension_flags field captures places where Code Judgment disagreed with Platform EM or where judgment is genuinely ambiguous — these become Google Doc items for strategic review, not code review.

post/api/v1/gm/code-judgment

Request

  • Base URL: https://app.goodleads.club (the document declares no server; this is the origin it was published from)
  • URL: https://app.goodleads.club/api/v1/gm/code-judgment
  • Auth: none declared

Request body

pr_numberinteger required
titlestring required
bodystring required
authorstring required
base_shastring required
head_shastring required
base_updated_atstring date-time required
files_changedstring[] required
additionsinteger required
deletionsinteger required
commit_messagesstring[] required
ci_conclusionstring required
ci_checksobject[] required
diffstring required
matrix_confidenceinteger nullable
platform_em_verdictobject nullable
labelsstring[]
behind_base_byinteger nullable

Response

Successful Response

unknown required

Changes

No recorded changes to this operation across all 1 revision of this API.