nisse-review: later reviews, replies, rounds, stop and review (M7.5)

Back on a PR it reviewed: a new head gets a later review of the commits
since (each one's diff, through the API) and of its earlier points; what
was said since gets replies, answered where it was said or skipped with a
reason. It signs off once and stays quiet while still happy. Past
tuning.review.max_rounds it sums up for the operator and goes quiet,
until a push or someone other than the PR's author joins in. An
operator's `stop` quiets it; `review` asks for a fresh review.

Reviews carry numbered markers and replies their point's key, so
recovery posts nothing twice.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01J5Utyfzw7gF7UhjRDEtjTk

docs/Architecture.md

@@ -902,7 +902,11 @@ else) towards better changes.
   what changed, judged by reading them; and the CI result as Gitea reports it.
 - **It discusses on the PR,** in inline threads and top-level comments, and
   answers replies to its points: it concedes a good argument and holds a point it
-  still believes, with reasons.
+  still believes, with reasons. Each thing said to it since its last look is
+  answered where it was said, or skipped with a reason (a bot's notice, say):
+  the model decides, as `nisse-dev` does (D5).
+- **Nothing to raise** is said once, as a short sign-off review. While it stays
+  happy, a later look that finds nothing new posts nothing.
 
 ### R2. Rounds
 
@@ -910,7 +914,8 @@ Two agents can argue forever, so a PR's exchanges are capped.
 
 - **A round** is one review or set of replies the reviewer posts. After a capped
   number of rounds (a tuning value) it posts a summary of what is still open, for
-  the operator, and stops answering on that PR.
+  the operator, and stops answering on that PR. Pushes have no cap of their own:
+  each brings new code (operator).
 - **The count resets** when someone other than the reviewer and the PR's agent
   author comments (in practice, the operator has joined the discussion), and when
   the PR gets a new push: new code is worth a fresh look.
@@ -955,8 +960,10 @@ Two agents can argue forever, so a PR's exchanges are capped.
   `ask_expert` for a frontier model's view (L9).
 - **No Matrix room** for now: the PR is its conversation. If every agent gets a
   room, it gets one too. Until then the operator stops it on a PR by commenting
-  `stop` there (a review is asked for again with `review`), and stops it entirely
-  by stopping its container.
+  `stop` there (a review is asked for again with `review`: a fresh one of the
+  whole PR, with the round count reset), and stops it entirely by stopping its
+  container. A command is an operator's whole top-level comment, acknowledged
+  with a reaction rather than a comment the PR's author would have to answer.
 - **Its own identity** (L12): a Gitea account (`nisse-review`) with read access to
   the repositories it reviews, its own token and LiteLLM key, and its own
   container on the agent host (X1), reaching only Gitea and LiteLLM.

docs/Development-Plan.md

@@ -660,9 +660,40 @@ a bot's comment needs a reply.
   inline too.
 - **Tests:** `test_loop_peers.py` over the fake Gitea.
 
-**M7.5 Discussion and rounds.** *Pre-approved.* Later reviews of what changed and
+**M7.5 Discussion and rounds.** *Done.* Later reviews of what changed and
 of earlier points, replies on threads, the round cap and its reset, `stop` and
-`review` on a PR (R1, R2, R5).
+`review` on a PR (R1, R2, R5):
+
+- **What wakes it:** each round, intake reads what was said on every open PR
+  (top-level comments, review bodies, line comments) and keeps the keys it has
+  taken in; anything new but its own is a point to answer. A run is due for a
+  PR never reviewed (or `review`ed), a new head (a **later review**, after its
+  CI as before), or new points at the same head (**replies**, no CI wait).
+- **A later review** is given the commits since its last review with each
+  one's diff (Gitea's compare lists them; the per-commit diff comes through the
+  API, since the web's compare diff takes no token), or the whole diff if they
+  can't be listed, and its earlier points with what was said on their lines.
+- **Replies:** `reply` answers a point where it was made (inline on its line,
+  else quoting it), `skip` leaves it with a reason, `done` ends; every point is
+  answered or skipped before a run may end. Each reply carries its point's key,
+  so a run re-done after a crash answers nothing twice.
+- **Sign-off:** `sign_off` posts a short review when there is nothing to raise;
+  while it stays happy, a later review offers `still_happy`, which posts nothing
+  (operator: no repeated LGTMs).
+- **Rounds:** each review, and each run that replied, is a round. Past
+  `tuning.review.max_rounds`, the next run sums up what is still open for the
+  operator (`post_summary`, as a review) and the item goes quiet (`capped`). A
+  push, or a comment from anyone but the reviewer and the PR's author (bots
+  included: the operator wants to see how the model handles them), resets it.
+- **`stop` and `review`:** an operator's whole comment; acknowledged with a 👍
+  reaction. `stop` quiets the item through pushes and comments; `review` asks
+  for a fresh review of the whole PR, and retries a failed item.
+- **Durability:** reviews carry a numbered marker (`<!-- nisse-review: {head}
+  #{n} -->`, the first in M7.3's form), so recovery asks Gitea about the right
+  one; items from before M7.5 take what was already said as heard.
+- **Tests:** `test_review_rounds.py`, including a no-dead-end walk over every
+  position; the live smoke test adds a push and a question, and checks the
+  later review at the new head.
 
 **M7.6 Deployment.** *Waits for the agent host (Phase 6).* Its image, container,
 Gitea account and secrets (R5).

docs/Tuning-Guide.md

@@ -637,7 +637,8 @@ reads file by file.
 ### `tuning.review.attempts` — default 2
 
 Failed review runs on one PR, at one head, before the reviewer says on the PR
-that it couldn't review it, and stops trying until the next push.
+that it couldn't review it, and stops trying until the next push or an
+operator's `review` comment.
 
 ### `tuning.review.ci_wait_s` — default 3600
 
@@ -661,3 +662,17 @@ review test, whose fresh repository has no CI).
 
 - **Too low:** a review starts before CI has reported, and runs beside it.
 - **Too high:** repositories without CI wait that long for every review.
+
+### `tuning.review.max_rounds` — default 4
+
+Rounds of discussion on one PR before the reviewer stops answering (R2): a
+round is a review it posts, or a run in which it replied. Past the cap, it
+posts one summary of what is still open, for the operator, and goes quiet
+on that PR. A new push, or a comment from anyone but the reviewer and the
+PR's author, starts the count again.
+
+- **Too low:** a disagreement worth one more exchange goes to the operator.
+- **Too high:** two agents argue at length (and GPU time) before anyone looks.
+- **Deciding:** a review, the author's answers, its replies, and one more
+  exchange is four; raise it if summaries arrive while the two are still
+  converging.