PR feedback

Author: wallstop

.devcontainer/devcontainer.json

@@ -146,4 +146,4 @@
 			}
 		}
 	}
-}
+}

.llm/design-history/README.md

@@ -13,4 +13,4 @@ Entry format:
 YYYY-MM-DD | Decision | Supersedes | Rationale | Evidence | PR/Issue
 ```
 
-Keep entries one line and append-only.
+Keep entries one line and append-only.

.llm/design-history/network.txt

@@ -1,2 +1,2 @@
 2025-12-15 | Validate frame numbers on all peer inputs | None | Prevent desync from malformed or malicious network packets | src/network/protocol | PR #200
-2026-03-01 | Prefer checked frame arithmetic in rollback paths | None | Prevent overflow-triggered desync at extreme frame values | src/sync_layer | PR #312
+2026-03-01 | Prefer checked frame arithmetic in rollback paths | None | Prevent overflow-triggered desync at extreme frame values | src/sync_layer | PR #312

.llm/design-history/sync-layer.txt

@@ -1 +1 @@
-2026-01-10 | Use fixed-capacity circular buffer for saved states | None | Keep rollback memory bounded and lookup cost stable | src/sync_layer/saved_states.rs | PR #280
+2026-01-10 | Use fixed-capacity circular buffer for saved states | None | Keep rollback memory bounded and lookup cost stable | src/sync_layer/saved_states.rs | PR #280

.llm/skills/workflows/adversarial-handoff.md

@@ -76,4 +76,4 @@ Exit Criteria:
 - Do not escalate every PR.
 - Do not run adversarial review without scoped questions.
 - Do not close handoff without regression coverage.
-- Do not merge High/Critical findings without explicit acceptance.
+- Do not merge High/Critical findings without explicit acceptance.

.llm/skills/workflows/design-decisions.md

@@ -71,4 +71,4 @@ During review:
 
 - Do not write essays. One line per decision.
 - Do not create logs for every small change.
-- Do not omit evidence links (file/symbol/PR).
+- Do not omit evidence links (file/symbol/PR).

.llm/skills/workflows/review-readiness.md

@@ -58,4 +58,4 @@ Review Readiness
 - Do not replace this checklist with prose summaries only.
 - Do not mark PASS without running the checks.
 - Do not treat this as a style checklist; this is correctness-first.
-- Do not open PRs with known blocking failures from this gate.
+- Do not open PRs with known blocking failures from this gate.

.llm/templates/ask-user-question.md

@@ -27,4 +27,4 @@ The missing detail is where the timeout occurs in the flow.
 - Does the timeout occur before handshake completion or during rollback replay?
 - Is it reproducible only under chaos settings (latency/loss), or also on clean UDP?
 
-If you do not have those details yet, I can start with a timeout-stage trace and narrow from first failing checkpoint.
+If you do not have those details yet, I can start with a timeout-stage trace and narrow from first failing checkpoint.