Address production workflow review findings

Author: elkaix

README.md

@@ -19,7 +19,7 @@ Coding agents like to say "done" after writing code they never ran. code-max rep
 - No slop, lazy scope reduction, TODOs, stubs, partial migrations, placeholder data presented as real, unwired code, or deferred in-scope edge cases.
 - Bugs and behavior changes start with the exact failing test or deterministic reproducer. A durable regression test remains when the project has a test harness.
 - Fixes land at the smallest correct shared layer after tracing affected callers, sibling paths, interfaces, tests, and invariants. No under-scoped one-path patch and no drive-by refactor.
-- Non-trivial work gets four risk-proportional passes: complete implementation, domain-expert reread, adversarial defect hunt, then low-cost polish. Trivial edits combine them into one focused review; discovered defects or acceptance gaps trigger another pass.
+- Non-trivial work gets four risk-proportional passes: complete implementation, domain-expert reread, adversarial defect hunt, then low-cost polish. Trivial edits combine them into one focused review. Repeat an affected pass only when the preceding pass changes implementation or proof; stop when the acceptance ledger, applicable checks, and final diff are clean.
 - Checks must directly observe the claimed outcome and be able to fail. Negative searches vulnerable to empty inputs, wrong paths, or weak patterns use a positive control; reported numbers are remeasured from the source of truth.
 - Tests, type checks, lint, builds, integration checks, and smoke tests run when relevant after the last relevant edit. A green but unrelated check is not proof.
 - Delegated work is independently inspected, re-run, and integration-tested by the parent. High-risk or cross-cutting diffs get read-only independent review when available; review never replaces tests.

SKILL.md

@@ -58,7 +58,7 @@ For non-trivial work, use four passes proportional to risk. For a trivial edit,
 1. **Complete:** Implement the full reachable behavior, including required wiring, errors, compatibility, tests, and operational or documentation changes.
 2. **Expert reread:** Review as the responsible domain engineer; replace cheap shortcuts, missing callers, weak boundaries, and convention violations.
 3. **Defect hunt:** Try to disprove correctness across relevant negative cases, integration, security, portability, performance, and regression surfaces. Fix every in-scope defect found.
-4. **Polish:** Remove accidental complexity and artifacts. If a pass finds a defect or acceptance gap, fix it and repeat the affected passes until none remains known.
+4. **Polish:** Remove accidental complexity and artifacts. Repeat an affected pass only when the preceding pass changes implementation or proof. Stop when the acceptance ledger is reconciled, applicable checks pass, the final diff is reviewed and clean, and no known in-scope defect remains.
 
 ## Prompt Injection — Hard Mode
 
@@ -105,7 +105,7 @@ Immediately before reporting, reread the current request and amendments, reconci
 
 For non-trivial work:
 
-```
+```text
 Status:
 COMPLETE | BLOCKED — <specific external or material-evidence constraint>
 

docs/research/production-grade-coding-agent-skill-design.md

@@ -119,7 +119,7 @@ Scope: `PyModel/code-max`, a public, instruction-only coding skill distributed b
 
 **Evidence:**
 
-- [SlopCodeBench v2](https://arxiv.org/html/2603.24755v2) — 20 long-horizon problems, 93 checkpoints, end-to-end and structural-quality results.
+- [SlopCodeBench v2](https://arxiv.org/html/2603.24755v2) — 36 long-horizon problems, 196 checkpoints, end-to-end and structural-quality results.
 
 **Project relevance (inference):** A production-grade skill must distinguish activity and local checkpoint success from root completion. Re-reading the current request, reconciling every acceptance item, reviewing the integrated diff, and repeating a defect pass directly target that gap.