[ai] Remove test code from release.

.agents/shared/test-loop.md

@@ -93,12 +93,18 @@ recovery is a repository hard stop.
 Impl commits must read like the repository's own history and carry only the durable task locator.
 Match the style of recent `git log` subjects.
 - **Subject:** one concise, plain-language line summarizing the change, ≤ ~50-60 characters. This is
-  the first line.
+  the first line. Start it with exactly `[ai] ` when the retained task implementation changes
+  permanent test-helper code, the agent harness, or agent documentation in any way; for every
+  other task, it must not contain `[ai]` anywhere. The prefix counts toward the length.
 - **Second line:** empty.
 - **Third line:** exactly `Task: <TASK_ID>`.
 - **Nothing else:** no explanatory body, `Autotask:`, attempt marker, `Co-Authored-By:`, or any
   tool/assistant attribution. The attempt number is runner state, never part of the message.
 
+The triggering scope includes `Telegram/SourceFiles/test/`, `.agents/`, `.claude/`, `AGENTS.md`,
+`CLAUDE.md`, and files whose sole role is supporting those systems. Classify the retained
+implementation only: the disposable test overlay and external AI task artifacts do not count.
+
 ## Test account (portable data) — hard rules
 
 The debug build runs in portable mode out of `out/Debug/`. Three sibling folders matter:

.agents/skills/perform-task/SKILL.md

@@ -89,7 +89,8 @@ task, canonical master must already contain its `Start` commit.
 Execute `references/pipeline.md` exactly. A normal task produces:
 
 1. one or more tested Telegram implementation-attempt commits, each with an
-   exact one-line subject, blank line, and `Task: <full-task-id>`;
+   exact one-line subject using the pipeline's conditional `[ai] ` prefix,
+   blank line, and `Task: <full-task-id>`;
 2. local tracked phase artifacts and progress in the AI slot worktree, without
    phase commits;
 3. one canonical `Approve <full-task-id>` commit containing all final AI

.agents/skills/perform-task/references/pipeline.md

@@ -277,7 +277,8 @@ Run sequentially:
    ```bash
    python3 SOURCE_ROOT/.agents/skills/process-inbox/scripts/workspace.py \
      source-commit --source-root SOURCE_ROOT --task TASK_ID \
-     --subject "<one concise plain-language subject>" --mark-green
+     --subject "<subject with the required conditional [ai] prefix>" \
+     --mark-green
    ```
 
    It verifies every dirty path against `work/owned-paths.txt` (plus the
@@ -376,11 +377,20 @@ unintended one can never slip into an attempt.
 Every implementation or implementation-fix commit message is exactly:
 
 ```text
-<one concise plain-language subject, about 50-60 characters>
+<conditional [ai] prefix><concise plain-language subject, about 50-60 characters>
 
 Task: <full TASK_ID>
 ```
 
+Determine the prefix from the retained task implementation, not from the
+temporary test overlay. Start the subject with exactly `[ai] ` when the task
+changes permanent test-helper code, the agent harness, or agent documentation
+in any way. This includes `Telegram/SourceFiles/test/`, `.agents/`, `.claude/`,
+`AGENTS.md`, `CLAUDE.md`, and files whose sole role is supporting those
+systems. The disposable test overlay and external AI task artifacts do not
+count. For every other task, the subject must not contain `[ai]` anywhere. The
+prefix counts toward the subject length.
+
 Do not add a body, `Autotask:`, attempt marker, `Co-Authored-By:`, assistant
 attribution, or any other trailer. Track rationale in the AI task. If a short
 durable explanation will help source-history readers, write