@jphutchins/code-review 0.1.0-alpha.45 → 0.1.0-alpha.47

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@jphutchins/code-review",
-  "version": "0.1.0-alpha.45",
+  "version": "0.1.0-alpha.47",
   "description": "Deterministic commenter for agentic PR review — gather, render, inline, post, adapt, extract, cost, validate, print-schema",
   "type": "module",
   "license": "MIT",

package/schema/VERSIONING.md

@@ -76,21 +76,25 @@ ignored); a version outside the supported set degrades to a §5.5 sticky notice.
 
 The commenter embeds the agent's **complete** findings document verbatim in review comments
 (`<!-- code-review:findings-json -->`) — no surfaced copy, no added fields (issue #156). The
-deterministic stop signal an iterating author-agent needs rides its own compact
-`<!-- code-review:signal -->` marker beside the blob: the pipeline-computed `convergence` (`{score,
-threshold, converged}` — a literal boolean, so a decoding agent cannot re-derive the weights) and
-`round` (the count of completed full-review rounds) of the last completed full-review round. The
-agent never writes the signal (it cannot know the score — the weights and threshold are
-commenter-side), so the findings schema above describes the whole embedded document; the signal
-marker declares its own surface version:
+deterministic stop signal an iterating author-agent needs is the pipeline-stamped `convergence` field
+INSIDE that document (issue #174): `{score, threshold, converged}` (a literal boolean, so a decoding
+agent cannot re-derive the weights) plus the per-round `rounds` trajectory. It is the single source of
+truth — the agent never writes it (it cannot know the score; the weights and threshold are
+commenter-side), and the findings schema above describes the whole embedded document, `convergence`
+included. When an oversized review falls to the link form (the blob is a URL, not base64), the same
+stamped object rides a compact `<!-- code-review:convergence -->` marker beside the link (issue #185)
+so the trajectory and stop signal survive; the embedded blob always wins on read.
+
+Two older surface formats survive only as READ-ONLY migration inputs — a re-review may still parse a
+sticky posted by an earlier release. Nothing writes either anymore:
 
 | Version | Status | Notes |
 |---|---|---|
-| `v0.7.0` | superseded | The pre-#156 surface axis: the commenter embedded a **surfaced** copy of the findings document carrying `convergence` + `round` inside it (issue #141). A sticky written by a `0.7.0` release still seeds — `stripSurfaceFields` peels the surfaced copy back to the agent's draft. |
-| `v0.8.0` | **current** | The version the compact `code-review:signal` marker declares (and the version a legacy `0.8.0` surfaced blob declared). `0.8.0` added the agent-facing `scope_metastasis` entry (issue #150): per-code consecutive-round recurrence counts plus a decision prompt. Post-#156 that entry is embedded in no document — the re-review seed re-derives it from the rounds marker and delivers it to the next-round agent. The flat draft schema (now `v0.9.0`) accepts an optional `scope_metastasis` property so a seed-echoing draft validates — an in-place additive change, deliberately NOT a draft version bump: a draft sharing a `0.7.0`/`0.8.0` number would collide with the surface axis's version gate, so the draft axis skips past them to `0.9.0` (issue #163); the axes must stay distinct so `stripSurfaceFields`/`parseSurfaceSignal` can tell a legacy surfaced blob from a draft). |
+| `v0.7.0` | legacy (read-only) | The pre-#156 surface axis: the commenter embedded a **surfaced** copy of the findings document carrying `convergence` + `round` inside it (issue #141). A sticky written by a `0.7.0` release still seeds — `stripSurfaceFields` peels the surfaced copy back to the agent's draft. |
+| `v0.8.0` | legacy (read-only) | The version a compact `<!-- code-review:signal -->` marker declared (and a legacy `0.8.0` surfaced blob). Its writer is retired (issue #186); `parseSignalMarker` / `parseSurfaceSignal` still decode it so a pre-#185 sticky seeds. `0.8.0` also added the agent-facing `scope_metastasis` entry (issue #150): per-code consecutive-round recurrence counts plus a decision prompt. Post-#156 that entry is embedded in no document — the re-review seed re-derives it from the carried trajectory. The flat draft schema (now `v0.9.0`) accepts an optional `scope_metastasis` property so a seed-echoing draft validates — an in-place additive change, deliberately NOT a draft version bump: a draft sharing a `0.7.0`/`0.8.0` number would collide with the surface axis's version gate, so the draft axis skips past them to `0.9.0` (issue #163); the axes must stay distinct so `stripSurfaceFields`/`parseSurfaceSignal` can tell a legacy surfaced blob from a draft). |
 
-The surface axis is independent of the draft-version registry: `v0.8.0` is the signal marker's
-contract, while the agent-written document above is now `v0.9.0`. `stripSurfaceFields` and
+The surface axis is independent of the draft-version registry: it numbers the legacy read-only signal
+formats, while the agent-written document above is `v0.9.0`. `stripSurfaceFields` and
 `parseSurfaceSignal` are version-gated on the surface axis, so a future draft bump can never be
 mistaken for a legacy surfaced blob — and no fresh document is ever surfaced.
 

package/schema/findings.schema.json

@@ -123,6 +123,55 @@
         }
       }
     },
+    "convergence": {
+      "type": "object",
+      "additionalProperties": false,
+      "description": "Advisory (issue #174): the current round's convergence score, the threshold it is judged against, whether it converged, and the per-round trajectory. The review agent does NOT write this field — the pipeline computes the score deterministically from the findings (floor(severity) + headroom × confidence × likelihood, summed) and stamps it, so this JSON document is the sole source of the convergence data a decoding agent reads. `score` and `threshold` are both carried so the number is interpretable on its own (converged when score <= threshold); prior rounds are carried verbatim, so changing the threshold never rewrites a past round's score.",
+      "required": ["score", "threshold", "converged"],
+      "properties": {
+        "score": {
+          "type": "number",
+          "description": "This round's convergence score — the sum of every finding's and systemic problem's contribution, rounded to 2 decimals. Lower is closer to convergence."
+        },
+        "threshold": {
+          "type": "number",
+          "description": "The convergence tolerance in force for this round: the round is converged when score <= threshold."
+        },
+        "converged": {
+          "type": "boolean",
+          "description": "Whether this round converged (score <= threshold). Advisory only — never alters the verdict."
+        },
+        "rounds": {
+          "type": "array",
+          "description": "The convergence trajectory: one entry per completed full-review round, oldest first, this round's entry last. Each score is a historical snapshot carried verbatim.",
+          "items": {
+            "type": "object",
+            "additionalProperties": false,
+            "required": ["round"],
+            "properties": {
+              "round": {
+                "type": "integer",
+                "minimum": 1,
+                "description": "The 1-indexed completed-round number this entry represents."
+              },
+              "score": {
+                "type": "number",
+                "description": "That round's convergence score, carried verbatim from when the round completed. Absent only for a round migrated from a pre-#174 sticky, which stored no score."
+              },
+              "codes": {
+                "type": "object",
+                "additionalProperties": { "type": "integer", "minimum": 0 },
+                "description": "That round's mechanism-frequency map: finding code -> number of findings carrying it. Feeds the same-mechanism and scope-metastasis recurrence signals."
+              },
+              "sha": {
+                "type": "string",
+                "description": "The reviewed head SHA (short form) for that round, so a same-head CI retry is recognised as the same iteration rather than new recurrence evidence."
+              }
+            }
+          }
+        }
+      }
+    },
     "findings": {
       "type": "array",
       "description": "Zero or more specific findings.",

package/templates/comment.eta

@@ -9,9 +9,6 @@
 <% if (it.findingsPointer) { -%>
 <%~ it.findingsPointer %>
 <% } -%>
-<% if (it.roundsMarker) { -%>
-<%~ it.roundsMarker %>
-<% } -%>
 
 ### <%= it.verdictBadge(it.findings.verdict) %>