name	description	user-invocable	allowed-tools
openstack-review	Review OpenStack Gerrit changes for code quality, correctness, and OpenStack best practices	true	Bash, Read, Grep, Glob, WebFetch, Write

OpenStack Code Review Skill

Review OpenStack Gerrit change $ARGUMENTS for code quality, compliance with OpenStack standards, and correctness.

Review Process

1. Fetch the Change

If not already checked out, fetch the review using the helper script:

# Use the helper script which reads project info from .gitreview
# Set SKILL_DIR based on where the skill is installed:
# - Project-specific: SKILL_DIR="./.claude/skills/openstack-review"
# - Global install: SKILL_DIR="$HOME/.claude/skills/openstack-review"
SKILL_DIR="${SKILL_DIR:-$HOME/.claude/skills/openstack-review}"

${SKILL_DIR}/fetch_review.sh $ARGUMENTS

Or use absolute path if you know the install location:

# If installed globally (most common)
~/.claude/skills/openstack-review/fetch_review.sh $ARGUMENTS

# If installed in project
./.claude/skills/openstack-review/fetch_review.sh $ARGUMENTS

Or manually:

# Extract project from .gitreview
PROJECT=$(grep "^project=" .gitreview | cut -d= -f2 | sed 's/\.git$//')
GERRIT_HOST=$(grep "^host=" .gitreview | cut -d= -f2)
PROJECT_ENCODED=$(echo "$PROJECT" | sed 's/\//%2F/g')

REVIEW_NUM=$ARGUMENTS
LAST_TWO=$(printf "%02d" $((REVIEW_NUM % 100)))

# Get latest patchset
PATCHSET=$(curl -s "https://${GERRIT_HOST}/changes/${PROJECT_ENCODED}~${REVIEW_NUM}/detail" | tail -n +2 | python3 -c "import sys, json; print(json.load(sys.stdin)['current_revision_number'])")

# Fetch and checkout
git fetch https://${GERRIT_HOST}/${PROJECT} refs/changes/${LAST_TWO}/${REVIEW_NUM}/${PATCHSET}
git checkout FETCH_HEAD

2. Fetch Previous Review Comments

Before analyzing the code, check what feedback other reviewers have already given:

Fetch review metadata and comments using the analyze_comments.py helper:

# Set SKILL_DIR based on where the skill is installed
SKILL_DIR="${SKILL_DIR:-$HOME/.claude/skills/openstack-review}"

# Extract project info from .gitreview
PROJECT=$(grep "^project=" .gitreview | cut -d= -f2 | sed 's/\.git$//')
GERRIT_HOST=$(grep "^host=" .gitreview | cut -d= -f2)
PROJECT_ENCODED=$(echo "$PROJECT" | sed 's/\//%2F/g')
REVIEW_NUM=$ARGUMENTS

# Fetch all review data including comments
curl -s "https://${GERRIT_HOST}/changes/${PROJECT_ENCODED}~${REVIEW_NUM}/detail?o=ALL_REVISIONS&o=ALL_COMMITS&o=ALL_FILES&o=MESSAGES&o=DETAILED_LABELS&o=DETAILED_ACCOUNTS" | tail -n +2 > /tmp/review_${REVIEW_NUM}.json

# IMPORTANT: The analyze_comments.py script expects a file path as argument
# Use the analyze_comments.py helper to parse and display review history
python3 ${SKILL_DIR}/analyze_comments.py /tmp/review_${REVIEW_NUM}.json

Note: The helper script will show:

Review metadata (subject, author, status, current patchset)
Patchset-by-patchset history of comments
Current review scores (Verified, Code-Review, Workflow)
Identification of potential issues (e.g., -1 votes)
Reviewers involved in the discussion

Fetch inline comments on specific files:

# Get comments for each patchset
for ps in $(seq 1 $(curl -s "https://${GERRIT_HOST}/changes/${PROJECT_ENCODED}~${REVIEW_NUM}/detail" | tail -n +2 | python3 -c "import sys, json; print(json.load(sys.stdin)['current_revision_number'])")); do
    echo "=== Patchset $ps Comments ==="
    curl -s "https://${GERRIT_HOST}/changes/${PROJECT_ENCODED}~${REVIEW_NUM}/revisions/$ps/comments" | tail -n +2 | python3 -m json.tool 2>/dev/null || echo "No inline comments"
done

3. Analyze the Changes

Run the following analysis:

Get commit details:

git show --stat
git show --format=fuller

List modified files:

git diff --name-only HEAD~1
git diff --stat HEAD~1

Review the actual changes:
```
git show HEAD
```

4. Verify Previous Feedback Has Been Addressed

CRITICAL: Review all previous comments from other reviewers:

Check each comment from previous patchsets:
- Read through all review messages
- Identify issues marked as needing fixes
- Look for -1 votes with specific concerns
- Note any unresolved discussions
Verify fixes in current patchset:
- For each issue raised, check if it's been addressed in the current code
- Look for changes in files that were previously criticized
- Verify test additions if tests were requested
- Check documentation updates if docs were requested
Common unaddressed issues to watch for:
- Reviewer asked for tests → Are they added?
- Reviewer asked for documentation → Is it updated?
- Reviewer pointed out security issue → Is it fixed?
- Reviewer suggested refactoring → Was it done?
- Reviewer asked questions → Are they answered in code or commit message?
If issues remain unaddressed:
- Note them in your review report
- Consider giving -1 with reference to the unaddressed feedback
- Ask the author to respond to previous comments

5. Code Quality Checks

Review the changes for OpenStack-specific requirements:

a. HACKING.rst Compliance

Check the project's HACKING.rst file for project-specific rules. Common OpenStack standards:

Import Order: Check that imports follow OpenStack ordering (stdlib, third-party, project)
Logging:
- Use LOG.warning() not LOG.warn()
- Don't translate log messages (no _() around logs)
- DO translate exception messages
JSON: Use oslo_serialization.jsonutils instead of json module
Line Continuation: No backslashes for line continuation (use parentheses)
Mutable Defaults: Method default arguments shouldn't be mutable
TaskFlow Reverts: TaskFlow revert methods must accept **kwargs (if project uses TaskFlow)

b. Test Coverage

Unit Tests Required: All new features must have unit tests
Bug Fix Tests: Bug fixes must include a test that fails without the fix
Test Structure: Test tree must mirror code structure
- Code: <project>/common/utils.py → Test: <project>/tests/unit/common/test_utils.py
Run structure validator if available: ./tools/check_unit_test_structure.sh
Check project's test requirements (often ≥92% coverage for Octavia, may vary by project)

c. Database Changes

Migration Required: Schema changes need an Alembic migration (location varies by project, typically <project>/db/migration/)
Repository Pattern: Check if project uses repository pattern for DB access
Model Updates: Changes to DB models must be compatible with existing data
Oslo.db: Verify proper use of oslo.db utilities

d. API Changes

Versioning: API changes must be versioned (check project's API structure)
Type Validation: Verify input validation (may use WSME, JSON Schema, or other frameworks)
Backward Compatibility: Don't break existing API contracts
Documentation: Update API reference docs
Microversions: Check if project uses API microversions

e. Configuration Changes

oslo.config: New config options need proper definitions
Sample Config: Run tox -e genconfig to update sample config
Documentation: Document new options in appropriate RST files

f. Security

Input Validation: All user inputs must be validated
SQL Injection: Use SQLAlchemy ORM, not raw SQL
Command Injection: Sanitize inputs to shell commands
Secrets: No hardcoded credentials or secrets
TLS/SSL: Check certificate validation is not disabled

6. Run Automated Checks

Execute the following tox environments:

Linting:
```
tox -e pep8
```
Unit Tests:
```
tox -e py3
```

Affected Tests: If you can identify specific affected tests, run them:

# CORRECT SYNTAX: Use -- to separate tox args from test args
tox -e py3 -- <project>.tests.unit.path.to.test_module

# Example for Octavia:
tox -e py3 -- octavia.tests.unit.api.common.test_types

Coverage Check:

tox -e cover
# Check project requirements (coverage thresholds vary by project)

7. Review Commit Message

Check that the commit message follows OpenStack format:

Short summary (50 chars max, no period)

Longer description explaining what and why (not how).
Wrap at 72 characters.

Can have multiple paragraphs.

Closes-Bug: #XXXXXX
Implements: blueprint name-of-blueprint
Change-Id: IXXXXXXXXXXXXXXXXXXXXXXXXXXX

Required elements:

Clear, descriptive summary
Detailed explanation in body
References to bugs/blueprints if applicable
Change-Id footer (added by git-review commit hook)
Signed-off-by if present

8. Generate Review Report

Create a markdown report with:

Summary: What the change does
Patchset History:
- Current patchset number
- Number of previous patchsets
- Previous review scores
Previous Review Feedback:
- Summary of comments from other reviewers
- Which issues have been addressed
- Which issues remain unresolved
- Any unanswered questions from reviewers
Files Changed: List of modified files
Code Quality Issues: Any HACKING.rst violations or style issues found
Testing:
- Are there tests?
- Do tests pass?
- Is coverage adequate?
- Were tests added per previous reviewer requests?
Security Concerns: Any security issues identified
Performance Impact: Any performance considerations
Backward Compatibility: Breaking changes?
Documentation: Is documentation updated/needed?
Response to Previous Feedback: Did the author address all previous comments?
Recommendation:

✅ Approve (+2/+1): Ready to merge, all feedback addressed
⚠️ Needs Work (-1): Issues to address (list specific items including unaddressed previous feedback)
❌ Reject (-2): Major problems

9. Additional Considerations

Design Principles: Check project-specific design principles (see project documentation)
- Is it idempotent?
- Is it scalable and resilient?
- Does it follow project conventions?
Dependencies: Any new dependencies added? Are they in requirements.txt?
- Check for conflicts with global-requirements
- Verify license compatibility
Documentation: Should this have a release note? (reno new slug-name)
Upgrade Impact: Does this affect rolling upgrades?
Configuration: New config options properly defined with oslo.config?

Example Usage

/openstack-review 970404

This will fetch review 970404, analyze it, run checks, and generate a comprehensive review report.

Output Format

The review should produce a detailed markdown report saved as REVIEW_${REVIEW_NUM}.md that can be used to:

Post feedback on Gerrit
Track review progress
Share with other reviewers
Document review decisions

Common Mistakes to Avoid

❌ Skill Path Issues

# WRONG - hardcoded path may not exist
python3 /some/fixed/path/analyze_comments.py /tmp/review_970404.json

# CORRECT - use SKILL_DIR variable for flexibility
SKILL_DIR="${SKILL_DIR:-$HOME/.claude/skills/openstack-review}"
python3 ${SKILL_DIR}/analyze_comments.py /tmp/review_970404.json

# OR use the appropriate path for your installation:
# Global install (most common):
python3 ~/.claude/skills/openstack-review/analyze_comments.py /tmp/review_970404.json

# Project-specific install:
python3 ./.claude/skills/openstack-review/analyze_comments.py /tmp/review_970404.json

❌ Wrong Filename for Comment Analysis Script

# WRONG - this file doesn't exist
python3 ${SKILL_DIR}/get_review_comments.py 970404

# CORRECT - the actual file is analyze_comments.py
SKILL_DIR="${SKILL_DIR:-$HOME/.claude/skills/openstack-review}"
python3 ${SKILL_DIR}/analyze_comments.py /tmp/review_970404.json

❌ Missing Review Data File for analyze_comments.py

# WRONG - script expects the data file as first argument, not review number
python3 ${SKILL_DIR}/analyze_comments.py 970404

# CORRECT - first fetch the data, then pass the file path
curl -s "https://review.opendev.org/changes/openstack%2Foctavia~970404/detail?o=ALL_REVISIONS&o=ALL_COMMITS&o=ALL_FILES&o=MESSAGES&o=DETAILED_LABELS&o=DETAILED_ACCOUNTS" | tail -n +2 > /tmp/review_970404.json
python3 ${SKILL_DIR}/analyze_comments.py /tmp/review_970404.json

❌ Incorrect tox Test Syntax

# WRONG - tox will error with "unrecognized arguments"
tox -e py3 octavia.tests.unit.api.common.test_types

# CORRECT - use -- to separate tox args from test args
tox -e py3 -- octavia.tests.unit.api.common.test_types

Quick Reference: Correct Usage Pattern

# 0. Set skill directory (adjust based on your installation)
# Default to global install, override if needed
SKILL_DIR="${SKILL_DIR:-$HOME/.claude/skills/openstack-review}"
# For project-specific install, use: SKILL_DIR="./.claude/skills/openstack-review"

# 1. Fetch the review
${SKILL_DIR}/fetch_review.sh 970404

# 2. Get review comments analysis
PROJECT=$(grep "^project=" .gitreview | cut -d= -f2 | sed 's/\.git$//')
GERRIT_HOST=$(grep "^host=" .gitreview | cut -d= -f2)
PROJECT_ENCODED=$(echo "$PROJECT" | sed 's/\//%2F/g')
REVIEW_NUM=970404

curl -s "https://${GERRIT_HOST}/changes/${PROJECT_ENCODED}~${REVIEW_NUM}/detail?o=ALL_REVISIONS&o=ALL_COMMITS&o=ALL_FILES&o=MESSAGES&o=DETAILED_LABELS&o=DETAILED_ACCOUNTS" | tail -n +2 > /tmp/review_${REVIEW_NUM}.json
python3 ${SKILL_DIR}/analyze_comments.py /tmp/review_${REVIEW_NUM}.json

# 3. Review the code
git show --stat
git show HEAD

# 4. Run automated checks
tox -e pep8
tox -e py3 -- <affected.test.module>

# 5. Generate review report
# Create REVIEW_${REVIEW_NUM}.md with findings

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

OpenStack Code Review Skill

Review Process

1. Fetch the Change

2. Fetch Previous Review Comments

3. Analyze the Changes

4. Verify Previous Feedback Has Been Addressed

5. Code Quality Checks

a. HACKING.rst Compliance

b. Test Coverage

c. Database Changes

d. API Changes

e. Configuration Changes

f. Security

6. Run Automated Checks

7. Review Commit Message

8. Generate Review Report

9. Additional Considerations

Example Usage

Output Format

Common Mistakes to Avoid

❌ Skill Path Issues

❌ Wrong Filename for Comment Analysis Script

❌ Missing Review Data File for analyze_comments.py

❌ Incorrect tox Test Syntax

Quick Reference: Correct Usage Pattern

FilesExpand file tree

SKILL.md

Latest commit

History

SKILL.md

File metadata and controls

OpenStack Code Review Skill

Review Process

1. Fetch the Change

2. Fetch Previous Review Comments

3. Analyze the Changes

4. Verify Previous Feedback Has Been Addressed

5. Code Quality Checks

a. HACKING.rst Compliance

b. Test Coverage

c. Database Changes

d. API Changes

e. Configuration Changes

f. Security

6. Run Automated Checks

7. Review Commit Message

8. Generate Review Report

9. Additional Considerations

Example Usage

Output Format

Common Mistakes to Avoid

❌ Skill Path Issues

❌ Wrong Filename for Comment Analysis Script

❌ Missing Review Data File for analyze_comments.py

❌ Incorrect tox Test Syntax

Quick Reference: Correct Usage Pattern