@@ -106,17 +106,15 @@ The workflow returns a review board grouped by outcome:
106106
107107- ` recommendedToPost ` — high-signal findings that look postable by a human
108108 reviewer and are not already covered
109- - ` possiblePlusOnes ` — existing threads where an endorsement may help
110- - ` partialOverlaps ` — findings that add useful detail beyond an existing thread
109+ - ` relatedToExisting ` — findings that overlap or endorse existing review threads
111110- ` discussionOnly ` — useful reviewer notes that should not be posted yet
112111- ` alreadyCovered ` — findings fully covered by existing human or bot review
113112 threads
114113- ` discarded ` — weak, duplicate, low-confidence, or non-actionable findings
115114
116115Each finding preserves the specialist's claim, evidence, reasoning, suggested
117116fix, confidence, source lens, and existing-review overlap rationale. The board
118- also includes positive observations, an action plan, coverage notes, PR metadata,
119- and review metadata.
117+ also includes positive observations, PR metadata, and review metadata.
120118
121119## Interaction And Posting
122120
@@ -212,7 +210,7 @@ tools, local-git runs record manifest/full-diff provenance in
212210` reviewMeta.sources ` , and MCP fallback runs record the fallback reason and
213211recovery counts. Also verify no generated parsing scripts are used, existing
214212review context affects recommendations, the review board is understandable, the
215- action plan is concise, drafts remain editable, and posting requires explicit
213+ drafts remain editable, and posting requires explicit
216214approval.
217215
218216## Prerequisites
0 commit comments