JSONbored · JSONbored · Jul 15, 2026 · Jul 15, 2026
@@ -19,7 +19,7 @@
 # --- Scope (per-repo cutover allowlist) -------------------------------------
 
 # Comma-separated owner/repo names allowed to run the per-PR review features
-# (SAFETY, GROUNDING, RAG, REPUTATION, UNIFIED_COMMENT). A per-PR feature runs on
+# (SAFETY, GROUNDING, RAG, REPUTATION). A per-PR feature runs on
 # a repo only if its own flag is ON *and* the repo is listed here. Empty = no
 # repos, so every per-PR feature stays dormant regardless of the flags below.
 # Case-insensitive, trimmed; stray commas ignored.
@@ -107,11 +107,6 @@ LOOPOVER_REVIEW_ENRICHMENT=false
 LOOPOVER_REVIEW_REPUTATION=false
 # GITTENSORY_REVIEW_REPUTATION=false                       # no longer read (removed by #4777)
 
-# Renders the public PR comment as one in-place unified comment instead of the
-# legacy multi-panel comment. OFF keeps the legacy comment byte-identical.
-LOOPOVER_REVIEW_UNIFIED_COMMENT=false
-# GITTENSORY_REVIEW_UNIFIED_COMMENT=false                       # no longer read (removed by #4777)
-
 # --- Global capabilities (NOT scoped by LOOPOVER_REVIEW_REPOS) -------------
 
 # Observability (read-only): cron anomaly scan over the gate-block ledger emits

@@ -71,8 +71,6 @@ REDIS_URL=redis://redis:6379
 SELFHOST_DEPLOYMENT_MODE=dry-run
 LOOPOVER_REVIEW_REPOS=owner/repo
 # GITTENSORY_REVIEW_REPOS=owner/repo        # no longer read (removed by #4777)
-LOOPOVER_REVIEW_UNIFIED_COMMENT=true
-# GITTENSORY_REVIEW_UNIFIED_COMMENT=true    # no longer read (removed by #4777)
 LOOPOVER_REVIEW_SAFETY=true
 # GITTENSORY_REVIEW_SAFETY=true             # no longer read (removed by #4777)
 LOOPOVER_REVIEW_GROUNDING=true

@@ -483,7 +483,6 @@ review:
     # max_files: 25
 
   # Unified-comment display toggles (#1957 / #1955 / #2069). Bool | null. Default: null/false — byte-identical.
-  # Only take effect when the `features.unifiedComment` feature is enabled.
   # changed_files_summary: false
   # When true, the unified review comment gains a deterministic "Changed files" summary table.
   # effort_score: false
@@ -1120,7 +1119,7 @@ settings:
   #   intentRouting: false # Closed-set intent classifier for unrecognized @loopover mentions -> existing Q&A commands only. Ollama-ONLY, same as chatQa (never uses chatQaFrontierFallback). Co-requisite: commandRateLimitPolicy: hold. Default: false.
 
 # Per-repo activation overrides for the converged review features that ship behind a deployment-wide
-# LOOPOVER_REVIEW_* env kill-switch (rag/reputation/unifiedComment/safety/grounding/e2eTests/screenshots/
+# LOOPOVER_REVIEW_* env kill-switch (rag/reputation/safety/grounding/e2eTests/screenshots/
 # improvementSignal). Each key is `true` (force on for this repo, subject to the env flag still being
 # enabled), `false` (force off), or omitted (falls back to the LOOPOVER_REVIEW_REPOS allowlist default --
 # an operator who sets nothing keeps today's behavior). Exception: `safety` is force-on-only -- an untrusted
@@ -1134,14 +1133,13 @@ settings:
 # `review.visual.enabled: false` as an always-available additional force-off); this key only answers "does
 # capture run for this repo at all," the same question its siblings answer for their own feature.
 # `e2eTests` (#4190) and `improvementSignal` (#4738, foundation phase of the #4737 PR-improvement-signal
-# epic) remain plain symmetric overrides like rag/reputation/unifiedComment -- neither carries a
+# epic) remain plain symmetric overrides like rag/reputation -- neither carries a
 # security-hardening or full-file-fetch rationale that would justify a force-on/force-off asymmetry.
 # `improvementSignal` is activation wiring only for now -- no tier reads the resolved
 # value yet.
 # features:
 #   rag: true
 #   reputation: false
-#   unifiedComment: true
 #   safety: true
 #   grounding: true
 #   e2eTests: false

@@ -21,7 +21,7 @@ The old vendor/embed plan is obsolete. The review system now lives in **gittenso
 
 - **Single project:** gittensory is the only source repo for the converged review system.
 - **Native port:** review features live under `src/review/**`, `src/queue/processors.ts`, and related first-party modules.
-- **Public comment path:** the unified in-place PR comment is driven by the native bridge and the `LOOPOVER_REVIEW_UNIFIED_COMMENT` flag.
+- **Public comment path:** the unified in-place PR comment is rendered unconditionally by the native bridge — it is the only comment-rendering path (the legacy multi-panel renderer and its `LOOPOVER_REVIEW_UNIFIED_COMMENT` flag were retired).
 - **Infra model:** D1 / Queue / AI / optional Vectorize / optional R2 / optional Browser bindings are declared directly in gittensory.
 - **Config model:** rollout is controlled by `LOOPOVER_REVIEW_*` flags plus the per-repo allowlist `LOOPOVER_REVIEW_REPOS`.
 - **Parity model:** parity is measured as a shadow/deploy-time comparison against authoritative legacy audit rows; local checkout validation proves structure and safety, not historical decision identity.
@@ -48,7 +48,6 @@ Those are operator actions. This repo should document them clearly and avoid imp
 
 Primary native review flags and surfaces:
 
-- `LOOPOVER_REVIEW_UNIFIED_COMMENT` — single public PR comment
 - `LOOPOVER_REVIEW_SAFETY` — prompt-injection defang + secret scan
 - `LOOPOVER_REVIEW_GROUNDING` — CI + full-file grounding
 - `LOOPOVER_REVIEW_RAG` — retrieval-augmented context

@@ -306,11 +306,6 @@ review:
           new / burst / low-reputation submitter to a deterministic-only review. Internal-only,
           never surfaced publicly. Per-PR.
         </li>
-        <li>
-          <code>LOOPOVER_REVIEW_UNIFIED_COMMENT</code> — renders the public PR comment as one
-          in-place unified comment instead of the legacy multi-panel comment. Per-PR; flag-off keeps
-          the legacy comment byte-identical.
-        </li>
         <li>
           <code>LOOPOVER_REVIEW_OPS</code> — read-only observability: a cron anomaly scan over your
           own review-outcome data plus a bearer-gated stats aggregate. Global (not scoped by the
@@ -338,9 +333,9 @@ review:
       </p>
       <CodeBlock
         lang="bash"
-        code={`# Roll grounding + the unified comment onto one repo:
+        code={`# Roll grounding + RAG context onto one repo:
 LOOPOVER_REVIEW_GROUNDING="true"
-LOOPOVER_REVIEW_UNIFIED_COMMENT="true"
+LOOPOVER_REVIEW_RAG="true"
 LOOPOVER_REVIEW_REPOS="JSONbored/loopover"`}
       />
 

@@ -230,8 +230,7 @@ function HowReviewsWork() {
       <h2>3. The unified review comment</h2>
       <p>
         The result is rendered as <strong>one in-place comment</strong> on the PR — updated in place
-        on each push rather than stacked — when <code>LOOPOVER_REVIEW_UNIFIED_COMMENT</code> is on
-        for the repo. It has three parts, top to bottom:
+        on each push rather than stacked. It has three parts, top to bottom:
       </p>
       <ul>
         <li>
@@ -240,14 +239,16 @@ function HowReviewsWork() {
           scans first.
         </li>
         <li>
-          <strong>The signal table</strong> — a compact row-per-signal summary: each dimension that
-          ran, its state (pass / advisory / block), and a short reason. This is the at-a-glance map
-          of why the verdict came out the way it did.
+          <strong>Decision drivers</strong> — a short, always-visible checklist of the signals that
+          actually decided the verdict: the AI code review and the gate result. This is the
+          at-a-glance answer to why the verdict came out the way it did.
         </li>
         <li>
-          <strong>Collapsibles</strong> — expandable sections for the detail behind each signal: the
-          AI review notes, the slop warnings, duplicate matches, manifest findings. Folded away by
-          default so the comment stays short, opened when a reader wants the evidence.
+          <strong>Collapsibles</strong> — expandable sections for everything else: a "Context &amp;
+          advisory signals" table (linked issue, related work, validation posture, contributor
+          context — signals that inform but never block the verdict), the AI review notes, slop
+          warnings, duplicate matches, manifest findings. Folded away by default so the comment
+          stays short, opened when a reader wants the evidence.
         </li>
       </ul>
       <p>
@@ -265,7 +266,10 @@ function HowReviewsWork() {
       </Callout>
 
       <h2>4. The signals behind a verdict</h2>
-      <p>Each row in the signal table comes from a named finding. The common ones you will see:</p>
+      <p>
+        Each row in the advisory signals table comes from a named finding. The common ones you will
+        see:
+      </p>
       <ul>
         <li>
           <code>secret_leak</code> — the safety scan (<code>LOOPOVER_REVIEW_SAFETY</code>) found a

@@ -89,7 +89,6 @@ LOOPOVER_REVIEW_IMPACT_MAP="true"              # deterministic impact map (needs
 LOOPOVER_REVIEW_CULTURE_PROFILE="true"         # repo quality-culture profile (needs review.culture_profile: true)
 LOOPOVER_REVIEW_MEMORY="true"                  # repeat-false-positive suppression (needs review.memory too)
 LOOPOVER_REVIEW_REPUTATION="true"              # submitter-reputation spend control (never shown)
-LOOPOVER_REVIEW_UNIFIED_COMMENT="true"         # one in-place unified PR comment
 LOOPOVER_REVIEW_ENRICHMENT="true"              # external analyzer registry (REES) findings
 LOOPOVER_REVIEW_INLINE_COMMENTS="true"         # diff-anchored inline PR review comments
 LOOPOVER_REVIEW_FIX_HANDOFF="true"             # machine-readable fix-handoff block (contributor-run)

@@ -435,7 +435,6 @@ GITHUB_METADATA_CACHE_TTL_SECONDS=600`}
       <CodeBlock
         filename=".env"
         code={`LOOPOVER_REVIEW_REPOS=owner/repo,owner/another
-LOOPOVER_REVIEW_UNIFIED_COMMENT=true
 LOOPOVER_REVIEW_INLINE_COMMENTS=false
 LOOPOVER_REVIEW_SAFETY=true
 LOOPOVER_REVIEW_GROUNDING=true
@@ -481,17 +480,16 @@ settings:
   agentDryRun: false
 features:
   safety: true
-  unifiedComment: true
   rag: false
   reputation: false`}
       />
       <p>
         The <code>features:</code> block above overrides a deployment-wide{" "}
-        <code>LOOPOVER_REVIEW_*</code> flag (rag, reputation, unifiedComment, safety) for this one
-        repo, with three states per key: <code>true</code> forces the capability on for this repo
-        (still subject to the env flag itself being enabled — it can never turn on a capability the
-        operator has fully disabled at the deployment level); <code>false</code> forces it off for
-        this repo regardless of the env flag; and omitting the key entirely falls back to the{" "}
+        <code>LOOPOVER_REVIEW_*</code> flag (rag, reputation, safety) for this one repo, with three
+        states per key: <code>true</code> forces the capability on for this repo (still subject to
+        the env flag itself being enabled — it can never turn on a capability the operator has fully
+        disabled at the deployment level); <code>false</code> forces it off for this repo regardless
+        of the env flag; and omitting the key entirely falls back to the{" "}
         <code>LOOPOVER_REVIEW_REPOS</code> allowlist default, i.e. today's behavior for an operator
         who hasn't set anything here. See <Link to="/docs/tuning">Tuning your reviews</Link> for the
         full <code>LOOPOVER_REVIEW_*</code> flag list this overrides.

@@ -194,11 +194,6 @@ function Tuning() {
           reputation proceeds normally. Never surfaced publicly — no comment, label, or check shows
           reputation. Per-PR.
         </li>
-        <li>
-          <code>LOOPOVER_REVIEW_UNIFIED_COMMENT</code> — renders the public PR comment as one
-          in-place unified comment instead of the legacy multi-panel comment. Per-PR. With the flag
-          off, the legacy comment is byte-identical.
-        </li>
         <li>
           <code>LOOPOVER_REVIEW_ENRICHMENT</code> — runs the review-enrichment analyzer registry
           (duplication, churn hotspots, blame links, approval integrity, undocumented exports, and

@@ -497,7 +497,6 @@ review:
     # max_files: 25
 
   # Unified-comment display toggles (#1957 / #1955 / #2069). Bool | null. Default: null/false — byte-identical.
-  # Only take effect when the `features.unifiedComment` feature is enabled.
   # changed_files_summary: false
   # When true, the unified review comment gains a deterministic "Changed files" summary table.
   # effort_score: false
@@ -1134,7 +1133,7 @@ settings:
   #   intentRouting: false # Closed-set intent classifier for unrecognized @loopover mentions -> existing Q&A commands only. Ollama-ONLY, same as chatQa (never uses chatQaFrontierFallback). Co-requisite: commandRateLimitPolicy: hold. Default: false.
 
 # Per-repo activation overrides for the converged review features that ship behind a deployment-wide
-# LOOPOVER_REVIEW_* env kill-switch (rag/reputation/unifiedComment/safety/grounding/e2eTests/screenshots/
+# LOOPOVER_REVIEW_* env kill-switch (rag/reputation/safety/grounding/e2eTests/screenshots/
 # improvementSignal). Each key is `true` (force on for this repo, subject to the env flag still being
 # enabled), `false` (force off), or omitted (falls back to the LOOPOVER_REVIEW_REPOS allowlist default --
 # an operator who sets nothing keeps today's behavior). Exception: `safety` is force-on-only -- an untrusted
@@ -1148,14 +1147,13 @@ settings:
 # `review.visual.enabled: false` as an always-available additional force-off); this key only answers "does
 # capture run for this repo at all," the same question its siblings answer for their own feature.
 # `e2eTests` (#4190) and `improvementSignal` (#4738, foundation phase of the #4737 PR-improvement-signal
-# epic) remain plain symmetric overrides like rag/reputation/unifiedComment -- neither carries a
+# epic) remain plain symmetric overrides like rag/reputation -- neither carries a
 # security-hardening or full-file-fetch rationale that would justify a force-on/force-off asymmetry.
 # `improvementSignal` is activation wiring only for now -- no tier reads the resolved
 # value yet.
 # features:
 #   rag: true
 #   reputation: false
-#   unifiedComment: true
 #   safety: true
 #   grounding: true
 #   e2eTests: false

@@ -261,7 +261,7 @@ export type CopycatGateMode = "off" | "warn" | "label" | "block";
 // level `false` behaves like any other plain override with no floor/ceiling. This is activation wiring only
 // -- no tier reads the resolved value yet (sibling sub-issues #4739-#4746 build the deterministic/LLM/panel
 // behavior that will gate on it).
-export const CONVERGED_FEATURE_KEYS = ["rag", "reputation", "unifiedComment", "safety", "grounding", "e2eTests", "screenshots", "improvementSignal"] as const;
+export const CONVERGED_FEATURE_KEYS = ["rag", "reputation", "safety", "grounding", "e2eTests", "screenshots", "improvementSignal"] as const;
 export type ConvergedFeatureKey = (typeof CONVERGED_FEATURE_KEYS)[number];
 
 /** Per-repo activation overrides for the converged review features (`features:` block). `true`/`false` force the
@@ -542,15 +542,15 @@ export type FocusManifestReviewConfig = {
    *  otherwise) — this is an ADDITIONAL opt-in on top of `review.inline_comments`, not a replacement gate.
    *  null/false (default, absent) = no suggestion blocks = byte-identical behavior. (#1956) */
   suggestions: boolean | null;
-  /** `review.changed_files_summary`: when true, the unified review comment (only rendered at all when the
-   *  `unifiedComment` convergence feature is on) gains a deterministic, no-AI "Changed files" collapsible: one
-   *  row per file category (source/test/docs/config/generated), with file counts and +/- totals, via the
-   *  existing `classifyChangedFile` classifier (`src/review/changed-files-classify.ts`, built for this table
-   *  under #2143). null/false (default, absent) = no changed-files section = byte-identical behavior. (#1957) */
+  /** `review.changed_files_summary`: when true, the unified review comment gains a deterministic, no-AI
+   *  "Changed files" collapsible: one row per file category (source/test/docs/config/generated), with file
+   *  counts and +/- totals, via the existing `classifyChangedFile` classifier
+   *  (`src/review/changed-files-classify.ts`, built for this table under #2143). null/false (default, absent)
+   *  = no changed-files section = byte-identical behavior. (#1957) */
   changedFilesSummary: boolean | null;
-  /** `review.effort_score`: when true, the unified review comment (only rendered when the `unifiedComment`
-   *  convergence feature is on) gains a compact "review effort: N/5 (~M min)" chip — a deterministic, no-AI
-   *  complexity/time estimate from `estimateReviewEffort` (`src/review/review-effort.ts`), weighting each
+  /** `review.effort_score`: when true, the unified review comment gains a compact "review effort: N/5 (~M min)"
+   *  chip — a deterministic, no-AI complexity/time estimate from `estimateReviewEffort`
+   *  (`src/review/review-effort.ts`), weighting each
    *  changed file's added lines by its category (source costs most; generated/vendored/lockfiles cost least)
    *  plus a fixed per-file overhead. Mirrors `changedFilesSummary` exactly: same table, same deterministic
    *  source, same display-only (never touches the AI prompt) shape. null/false (default, absent) = no chip =
@@ -1021,7 +1021,6 @@ const EMPTY_FEATURES_CONFIG: FocusManifestFeaturesConfig = {
   present: false,
   rag: null,
   reputation: null,
-  unifiedComment: null,
   safety: null,
   grounding: null,
   e2eTests: null,

@@ -302,11 +302,6 @@ declare global {
      *  secret, never a public var. When absent, BYOK is unavailable and review uses the configured instance
      *  reviewer when available. */
     TOKEN_ENCRYPTION_SECRET?: string;
-    /** #6103: retired -- the unified-comment bridge is now the ONLY PR-comment renderer unconditionally (the
-     *  legacy `buildPublicPrIntelligenceComment` panel this flag used to switch away from was deleted, having
-     *  no remaining production caller). No longer read; kept only so an operator's existing deployment config
-     *  setting it doesn't error. Safe to remove from any env once noticed. */
-    LOOPOVER_REVIEW_UNIFIED_COMMENT?: string;
     /** Inline comments (#inline-comments): when truthy (AND the repo is in LOOPOVER_REVIEW_REPOS AND the repo's
      *  `.loopover.yml` sets `review.inline_comments: true`), the AI reviewer ALSO leaves quiet, NON-BLOCKING
      *  inline comments on specific changed lines, layered on top of the decision summary. Default OFF —

@@ -348,8 +348,8 @@ export function refreshLiveMergeState(
 // refreshLiveMergeState's #4220 doc comment above prohibits for this act-boundary-adjacent disposition input.
 // So this only reuses a memoized value when its key is ALSO in forcedMergeStateKeys/forcedCiAggregateKeys --
 // i.e. it was written by a FORCED (genuinely-live-this-pass) call, never by a cached-path reader. On a genuine
-// miss (no forced call ran yet this pass, e.g. unifiedCommentAllowed was false) this falls through to a REAL
-// live refresh, so behavior can only ever improve (fewer calls) over the pre-fix code, never go staler.
+// miss (no forced call has run yet this pass) this falls through to a REAL live refresh, so behavior can only
+// ever improve (fewer calls) over the pre-fix code, never go staler.
 export function reuseOrRefreshLiveMergeState(
   env: Env,
   repoFullName: string,

@@ -1,7 +1,7 @@
 // Convergence (cutover) per-repo gate: an allowlist that activates the PER-PR converged review features one
 // repo at a time, so the cutover can be rolled forward (and rolled back) on a single repo without flipping the
 // global flags off for everyone. Each per-PR converged feature ALREADY has a global switch (LOOPOVER_REVIEW_SAFETY /
-// _GROUNDING / _RAG / _REPUTATION, LOOPOVER_REVIEW_UNIFIED_COMMENT); this adds a SECOND, repo-scoped gate that must
+// _GROUNDING / _RAG / _REPUTATION); this adds a SECOND, repo-scoped gate that must
 // ALSO pass for the feature to run on a given PR's repo.
 //
 // Single env var: LOOPOVER_REVIEW_REPOS — a comma-separated allowlist of repo full-names