style(release): drop the verbose explanatory comments

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Author: yanxue06

.github/blocks/bump-monorepo-versions/action.yaml

@@ -96,14 +96,8 @@ runs:
           CURRENT_VER=$(jq -r '.version // "0.0.0"' "$PKG_PATH/package.json")
 
           COMMITS=$(git log --oneline "${LAST_SHA}..HEAD" -- "$PKG_PATH" 2>/dev/null || git log --oneline HEAD~10..HEAD -- "$PKG_PATH")
-          # Pre-compute the diff per package and inject it — codex-action does not
-          # auto-attach a diff, so the version must be decided on the actual code,
-          # not just commit messages (see determine-publish-version for the why).
           NAME_STATUS=$(git diff --name-status "${LAST_SHA}..HEAD" -- "$PKG_PATH" 2>/dev/null || true)
           PKG_DIFF=$(git diff "${LAST_SHA}..HEAD" -- "$PKG_PATH" 2>/dev/null || true)
-          # Per-package cap: the prompt output concatenates every changed
-          # package's diff into one step output, which GitHub caps at 1 MB/job —
-          # keep headroom for several changed packages.
           CAP=100000
           if [ "${#PKG_DIFF}" -gt "$CAP" ]; then
             PKG_DIFF="${PKG_DIFF:0:$CAP}
@@ -134,9 +128,6 @@ runs:
       with:
         openai-api-key: ${{ inputs.openai-api-key }}
         safety-strategy: read-only
-        # Diff-based breaking-change judgment is a real reasoning task — run at
-        # high effort. The diff is provided per package below; the agent may also
-        # run read-only git for any truncated file.
         effort: high
         prompt: |
           You are a release manager for the "${{ inputs.service-name }}" monorepo.

.github/blocks/determine-publish-version/action.yaml

@@ -84,20 +84,10 @@ runs:
         HEAD_SHA: ${{ github.sha }}
       run: |
         set -uo pipefail
-        # Pre-compute the diff and inject it into the prompt. openai/codex-action
-        # does NOT auto-attach a diff — it is an agentic `codex exec`, so left to
-        # itself the agent may only run `git log` (commit messages) and never look
-        # at the code (exactly what happened on the v0.2.0-rc.11 run). Embedding
-        # the diff here GUARANTEES the version is decided on the actual changes.
-        # name-status + churn are always complete (they reveal removed/renamed
-        # files — the strongest breaking signal); the full diff is capped to keep
-        # the prompt bounded, and the agent can pull any truncated file via git.
         NAME_STATUS=$(git diff --name-status "$PREV_SHA" "$HEAD_SHA" 2>/dev/null || true)
         STAT=$(git diff --stat "$PREV_SHA" "$HEAD_SHA" 2>/dev/null || true)
         LOG=$(git log --no-merges --format='- %s%n%b' "$PREV_SHA..$HEAD_SHA" 2>/dev/null || true)
         DIFF=$(git diff "$PREV_SHA" "$HEAD_SHA" 2>/dev/null || true)
-        # The diff is injected via a step output, which GitHub caps at 1 MB/job;
-        # 400 KB leaves headroom for the log + stat. Don't raise past ~800 KB.
         CAP=400000
         TRUNC=""
         if [ "${#DIFF}" -gt "$CAP" ]; then
@@ -130,9 +120,6 @@ runs:
       with:
         openai-api-key: ${{ inputs.openai-api-key }}
         safety-strategy: read-only
-        # Judging backward-compatibility from a diff is a real reasoning task —
-        # run it at high effort. This step runs once per release and a wrong
-        # major bump is expensive, so the budget is justified.
         effort: high
         prompt: |
           Determine the next semantic version after ${{ steps.last-release.outputs.version }}.