refactor(skills): переработан плейбук координации Codex

- Зачем:
  - Актуализирована координация под Astra с сохранением замысла и бюджета контекста.
- Что:
  - Разделены роли Astra, Sol и Luna; выбор приемов оставлен координатору.
  - Сохранены независимое ревью, триаж и проверка соразмерности исправлений.
  - Обновлены нативные механики и необязательная линия Claude CLI.
- Проверка:
  - Проверены Markdown-ссылки, YAML и git diff --check; описано ограничение валидатора.
  - Пройден поведенческий сценарий с тремя тестами и независимое ревью Claude.
  - Перепроверены находки и фактическое продолжение фоновой сессии Claude.

agents/.agents/skills/codex-subagent-playbook/references/prompts/review-boilerplate.md

@@ -1,40 +1,44 @@
-## Mandate
+## Reviewer mandate
 
-- Review independently and test hypotheses. Do not edit source files,
-  persistent environment state, or Git.
-- Use the working permissions needed for checks, caches, temporary artifacts,
-  documentation, and primary-source research.
-- Write the detailed report only to REPORT. This assigned exchange file is the
-  sole write exception to the observer mandate.
-- Treat another agent's claims as untrusted until evidence supports them.
+- Review the stated version independently. Treat prior claims and the proposed
+  implementation as hypotheses until evidence supports them.
+- Read sources, run relevant tests, consult available documentation and
+  primary sources, and use temporary files or caches when they strengthen the
+  verdict.
+- Preserve source files, Git, services, data, and other persistent state.
+  Write the detailed result only to REPORT.
+- Judge defects at the reviewed artifact's own level and price a proposed fix
+  against the risk it removes. A finding is a claim, not an order.
 
 ## Detailed report
 
-Use stable IDs with FINDING PREFIX:
+Use permanent IDs with FINDING PREFIX.
 
-`PREFIX-N | BLOCKER|MAJOR|MINOR|NIT | file:line or task section | violated
-criterion | failure scenario | evidence | minimal verification`
+`PREFIX-N | location | violated criterion |
+failure scenario | evidence | minimal verification`
 
-A non-gating consideration has no stable ID, severity, or counter. Use it only
-for a material risk from a safely deferred decision, and do not prescribe the
-downstream implementation. It does not change the verdict.
+When the project or review line already defines a severity scale, add severity
+to the record and count it.
 
-Use CHANGES_REQUESTED only when at least one finding requires revision at the
-reviewed artifact's own level. APPROVED means that no such defect remains; it
-does not mean that every downstream question has been answered.
+A non-gating consideration has no ID, severity, or counter. Use it only for a
+material risk of a safely deferred decision. It does not change the verdict.
+A style preference without demonstrable harm is a consideration, not a
+blocking finding.
 
-If there are no findings, record the reviewed scope and every executed check
-or tested hypothesis with its outcome. A bare approval is invalid.
+`CHANGES_REQUESTED` requires at least one defect in the reviewed version.
+`APPROVED` means no such defect remains. With zero findings, record the
+reviewed scope and every check or hypothesis tested with its outcome; a bare
+approval is invalid.
 
 ## Final message
 
 Return only:
 
 - `VERDICT: APPROVED | CHANGES_REQUESTED`
-- severity counters;
-- one line per finding: `ID | severity | area | one-phrase essence`;
-- optional one-line material considerations:
-  `CONSIDERATION | area | one-phrase risk`;
-- `REPORT: <absolute path>`.
+- available counters
+- one line per finding with ID, area, one-phrase claim, and any established
+  severity
+- every material consideration, one line each
+- `REPORT: <absolute path>`
 
-Do not repeat the detailed report or reasoning narrative in the final message.
+Do not repeat the detailed report.