prr-kit 2.0.10 → 2.0.12

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "prr-kit",
-  "version": "2.0.10",
+  "version": "2.0.12",
   "description": "AI-driven Pull Request Review Kit — structured agent workflows for thorough, consistent code review",
   "main": "tools/cli/prr-cli.js",
   "bin": {

package/src/core/tasks/workflow.xml

@@ -7,6 +7,7 @@
     <rule>ALWAYS halt at decision points and wait for user input</rule>
     <rule>ALWAYS load config variables before starting any workflow</rule>
     <rule>Use native file-reading tools for file content — avoid shell commands for reading files</rule>
+    <rule>For runtime code execution (file parsing, JSON building, temp file operations), use Node.js — do not assume Python is available</rule>
   </rules>
 
   <variable-resolution>

package/src/prr/workflows/6-report/post-comments/steps/step-01-format.md

@@ -149,22 +149,14 @@ Bash heredoc breaks with Unicode characters, emojis, backticks, and nested quote
 2. **Execute** it with `node`
 3. Script writes the final JSON to `{temp_dir}/prr-payload.json`
 
-**Step A — Detect available runtime:**
-```bash
-node --version 2>/dev/null && echo "use-node" || (python3 --version 2>/dev/null && echo "use-python3") || echo "no-runtime"
-```
-Prefer `node`. Fallback: `python3`, then `python`.
+**Step A — Write the build script using the Write tool** (NOT echo/heredoc/Bash):
 
-**Step B — Write the build script using the Write tool** (NOT echo/heredoc/Bash):
-
-Use the Write tool to create `{temp_dir}/build-payload.mjs` (Node.js) or `{temp_dir}/build-payload.py` (Python).
+Use the Write tool to create `{temp_dir}/build-payload.mjs`.
 
 The script must:
-- Define all string values (comment bodies, summary) as native string variables — no manual JSON escaping needed, the runtime handles it
+- Define all string values (comment bodies, summary) as native string variables — no manual JSON escaping needed, Node.js handles it
 - Build the payload object in memory
-- Write to `{temp_dir}/prr-payload.json` using native JSON serialization:
-  - Node.js: `JSON.stringify(payload, null, 2)`
-  - Python: `json.dumps(payload, ensure_ascii=False, indent=2)`
+- Write to `{temp_dir}/prr-payload.json` using `JSON.stringify(payload, null, 2)`
 
 **Node.js template** (`build-payload.mjs`):
 ```js
@@ -184,36 +176,14 @@ writeFileSync("{temp_dir}/prr-payload.json", JSON.stringify(payload, null, 2), "
 console.log(`OK: ${payload.comments.length} comments`)
 ```
 
-**Python template** (`build-payload.py`):
-```python
-import json
-
-payload = {
-    "commit_id": "COMMIT_SHA",
-    "body": "SUMMARY BODY",
-    "event": "REQUEST_CHANGES",
-    "comments": [
-        {"path": "src/file.js", "line": 42, "side": "RIGHT", "body": "🔴 **[BLOCKER]** ..."},
-    ]
-}
-
-with open("{temp_dir}/prr-payload.json", "w", encoding="utf-8") as f:
-    json.dump(payload, f, ensure_ascii=False, indent=2)
-print(f"OK: {len(payload['comments'])} comments")
-```
-
-**Step C — Execute:**
+**Step B — Execute:**
 ```bash
 node "{temp_dir}/build-payload.mjs"
-# or: python3 "{temp_dir}/build-payload.py"
 ```
 
-**Step D — Verify output (optional sanity check):**
+**Step C — Verify output (optional sanity check):**
 ```bash
-# Node.js
 node -e "const p=require('{temp_dir}/prr-payload.json'); console.log('OK:', p.comments.length, 'comments')"
-# Python
-python3 -c "import json; p=json.load(open('{temp_dir}/prr-payload.json')); print('OK:', len(p['comments']), 'comments')"
 ```
 
 ---

package/src/prr/workflows/6-report/post-comments/steps/step-02-post.md

@@ -27,14 +27,9 @@ description: "Post comments via platform CLI/API, verify, clean up"
 **Case A — Self-review** (`"Review Can not request changes on your own pull request"`):
 - The authenticated user is the PR author. GitHub disallows `REQUEST_CHANGES` on own PRs.
 - Fix: Update `event` in the payload from `"REQUEST_CHANGES"` to `"COMMENT"` and retry once.
-- Use whichever runtime is available (`{detected_runtime}` from step A earlier):
 
 ```bash
-# Node.js
 node -e "const fs=require('fs'); const p=JSON.parse(fs.readFileSync('{temp_dir}/prr-payload.json','utf-8')); p.event='COMMENT'; fs.writeFileSync('{temp_dir}/prr-payload.json',JSON.stringify(p,null,2)); console.log('event → COMMENT')"
-
-# Python3
-python3 -c "import json; f=open('{temp_dir}/prr-payload.json','r+'); p=json.load(f); p['event']='COMMENT'; f.seek(0); json.dump(p,f,ensure_ascii=False,indent=2); f.truncate(); print('event → COMMENT')"
 ```
 Then retry the POST above.
 
@@ -79,7 +74,7 @@ done
 
 Use the runtime script approach to build the summary payload first:
 
-Write `{temp_dir}/build-bb-summary.mjs` (or `.py`) using the Write tool:
+Write `{temp_dir}/build-bb-summary.mjs` using the Write tool:
 
 ```js
 // Node.js
@@ -167,9 +162,7 @@ rm -f "{temp_dir}/prr-payload.json" \
        "{temp_dir}/prr-bb-"*.json \
        "{temp_dir}/prr-bb-summary.json" \
        "{temp_dir}/build-payload.mjs" \
-       "{temp_dir}/build-payload.py" \
-       "{temp_dir}/build-bb-summary.mjs" \
-       "{temp_dir}/build-bb-summary.py"
+       "{temp_dir}/build-bb-summary.mjs"
 rmdir "{temp_dir}" 2>/dev/null || true
 ```
 

package/src/prr/workflows/quick/workflow.md

@@ -180,28 +180,7 @@ Wait for response. If empty → detect `origin/main` or `origin/master`.
 Set `base_branch` = input or detected default.
 Set `diff_range` = `{base_branch}...{target_branch}`.
 
-### 1d. Load diff
-
-Use the first available method based on platform:
-
-**GitHub** (if `active_platform = github` and `pr_number` is set):
-```bash
-gh pr diff {pr_number} --repo {active_platform_repo}
-```
-
-**GitLab** (if `active_platform = gitlab` and `pr_number` is set):
-```bash
-glab mr diff {pr_number} --repo {active_platform_repo}
-```
-
-**Azure DevOps / Bitbucket / fallback:**
-```bash
-git -C {target_repo} diff {base_branch}...{target_branch} --stat
-git -C {target_repo} diff {base_branch}...{target_branch}
-```
-Store diff in memory. Count files changed, lines added/removed.
-
-### 1e. Create Session Folder
+### 1d. Create Session Folder
 
 Compute the session output folder:
 
@@ -233,9 +212,30 @@ mkdir -p "{session_output}"
 
 **Store in working context:** `session_output`, `target_branch`, `base_branch`, `pr_number`, `pr_title`, `active_platform`, `active_platform_repo`.
 
+### 1e. Load diff
+
+Use the first available method based on platform:
+
+**GitHub** (if `active_platform = github` and `pr_number` is set):
+```bash
+gh pr diff {pr_number} --repo {active_platform_repo} > "{session_output}/full.diff"
+```
+
+**GitLab** (if `active_platform = gitlab` and `pr_number` is set):
+```bash
+glab mr diff {pr_number} --repo {active_platform_repo} > "{session_output}/full.diff"
+```
+
+**Azure DevOps / Bitbucket / fallback:**
+```bash
+git -C {target_repo} diff {base_branch}...{target_branch} > "{session_output}/full.diff"
+```
+
+Save the full diff to `{session_output}/full.diff`. Read that file to count files changed, lines added/removed.
+
 ### 1f. Generate Diffs Folder
 
-Parse the diff loaded in 1d and write per-file markdown files under `{session_output}/diffs/`, mirroring the repo folder tree.
+Parse `{session_output}/full.diff` and write per-file markdown files under `{session_output}/diffs/`, mirroring the repo folder tree.
 
 **For each changed file in the diff:**
 
@@ -275,6 +275,8 @@ After writing all files, show summary:
 ✓ Diffs saved: {file_count} files → {session_output}/diffs/
 ```
 
+**⛔ After this point, do NOT run any further `git diff` commands to read file changes. All diff content is now available at `{session_output}/diffs/`. Read those files directly for all subsequent phases.**
+
 ---
 
 ## PHASE 2 — DESCRIBE PR