Skip to content

feat: integrate stillriver-ai-workflows action #25

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
adslaton merged 5 commits into from
Jul 29, 2025
Merged

feat: integrate stillriver-ai-workflows action #25

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
0a9b74a
feat: integrate stillriver-ai-workflows action
adslaton Jul 29, 2025
b3d18a7
fix: use environment variables for action outputs in github-script
adslaton Jul 29, 2025
12f257a
feat: add rate limiting and inline comments configuration
adslaton Jul 29, 2025
98413aa
fix: improve env variable scoping in workflow
adslaton Jul 29, 2025
b90d26d
feat: enable GitHub native resolvable comments
adslaton Jul 29, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
243 changes: 34 additions & 209 deletions .github/workflows/ai-pr-review.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -72,6 +72,8 @@ jobs:
- name: Check for recent AI review
id: check_recent_review
uses: actions/github-script@v7
env:
RATE_LIMIT_MINUTES: ${{ vars.AI_REVIEW_RATE_LIMIT_MINUTES || '1' }}
with:
github-token: ${{ github.token }}
script: |
Expand All @@ -82,16 +84,18 @@ jobs:
per_page: 100
});

// Check if there's a recent AI review (within last 5 minutes)
const fiveMinutesAgo = new Date(Date.now() - 5 * 60 * 1000);
// Check if there's a recent AI review (within rate limit window)
const rateLimitMinutes = parseInt(process.env.RATE_LIMIT_MINUTES) || 1;
const rateLimitMs = rateLimitMinutes * 60 * 1000;
const rateLimitAgo = new Date(Date.now() - rateLimitMs);
const recentAIReview = comments.data.find(comment =>
comment.user.login === 'github-actions[bot]' &&
comment.body.includes('🤖 AI Review by Claude') &&
new Date(comment.created_at) > fiveMinutesAgo
comment.body.includes('🤖 AI Review') &&
new Date(comment.created_at) > rateLimitAgo
);

if (recentAIReview) {
console.log('Recent AI review found, skipping to prevent spam');
console.log(`Recent AI review found within ${rateLimitMinutes} minute(s), skipping to prevent spam`);
core.setOutput('skip', 'true');
} else {
core.setOutput('skip', 'false');
Expand Down Expand Up @@ -166,215 +170,28 @@ jobs:
if: (steps.check_tests.outputs.result == 'true' || github.event_name != 'pull_request') && steps.check_recent_review.outputs.skip != 'true'
uses: actions/checkout@v4
with:
# Fetch the PR branch - use get_pr outputs for comment triggers
ref: ${{ github.event.pull_request.head.sha || steps.get_pr.outputs.head_sha || github.event.pull_request.head.ref || steps.get_pr.outputs.head_ref }}
fetch-depth: 0

- name: Setup Python
if: (steps.check_tests.outputs.result == 'true' || github.event_name != 'pull_request') && steps.check_recent_review.outputs.skip != 'true'
uses: actions/setup-python@v5
with:
python-version: '3.11'

- name: Install OpenRouter dependencies
if: (steps.check_tests.outputs.result == 'true' || github.event_name != 'pull_request') && steps.check_recent_review.outputs.skip != 'true'
run: |
pip install openai==1.54.3 httpx==0.27.0

- name: Get PR diff
if: (steps.check_tests.outputs.result == 'true' || github.event_name != 'pull_request') && steps.check_recent_review.outputs.skip != 'true'
id: diff
run: |
# Get the base branch - use get_pr outputs for comment triggers
BASE_BRANCH="${{ github.event.pull_request.base.ref || steps.get_pr.outputs.base_ref || 'main' }}"

# Get the diff
git diff origin/${BASE_BRANCH}...HEAD > pr_diff.txt

# Also get list of changed files
git diff --name-only origin/${BASE_BRANCH}...HEAD > changed_files.txt

# Get PR description - use safe method to avoid shell interpretation
cat > pr_description.txt << 'EOF'
${{ github.event.pull_request.body || steps.get_pr.outputs.body || '' }}
EOF

- name: Validate input files
if: (steps.check_tests.outputs.result == 'true' || github.event_name != 'pull_request') && steps.check_recent_review.outputs.skip != 'true'
run: |
# Check if required files exist
if [ ! -f pr_diff.txt ]; then
echo "Error: pr_diff.txt not found"
exit 1
fi
if [ ! -f changed_files.txt ]; then
echo "Error: changed_files.txt not found"
exit 1
fi
if [ ! -f pr_description.txt ]; then
echo "Error: pr_description.txt not found"
exit 1
fi
echo "All required files found"

- name: Run AI Review via OpenRouter
- name: Run AI Review using stillriver-ai-workflows
if: (steps.check_tests.outputs.result == 'true' || github.event_name != 'pull_request') && steps.check_recent_review.outputs.skip != 'true'
id: ai_review
uses: stillrivercode/stillriver-ai-workflows@v1
env:
OPENROUTER_API_KEY: ${{ secrets.OPENROUTER_API_KEY }} # pragma: allowlist secret
AI_MODEL: ${{ vars.AI_MODEL || 'anthropic/claude-sonnet-4' }}
run: | # pragma: allowlist secret
# Validate API key exists # pragma: allowlist secret
if [[ -z "$OPENROUTER_API_KEY" ]]; then # pragma: allowlist secret
echo "## ⚠️ AI Review Failed" > review_output.md
echo "OPENROUTER_API_KEY not configured in repository secrets." >> review_output.md
exit 0
fi

# Create review prompt
cat > review_prompt.txt << EOF
Please review this pull request and provide feedback. Focus on:
1. Code quality and best practices
2. Potential bugs or issues
3. Security concerns
4. Performance implications
5. Test coverage
6. Documentation updates needed

PR Description:
$(cat pr_description.txt)

Changed files:
$(cat changed_files.txt)

Diff (truncated to first 5000 lines if longer):
$(cat pr_diff.txt | head -5000)

Please provide a structured review with:
- Summary of changes
- Strengths of the implementation
- Issues or concerns (if any)
- Suggestions for improvement
- Overall recommendation (approve, request changes, or comment)
EOF

# Create OpenRouter API script
cat > ai_review.py << 'EOF'
import os
import sys
import json
from openai import OpenAI

def main():
api_key = os.environ.get('OPENROUTER_API_KEY')
model = os.environ.get('AI_MODEL', 'anthropic/claude-sonnet-4')

if not api_key:
print("## ⚠️ AI Review Failed")
print("OPENROUTER_API_KEY not configured.")
return 1

# Read the prompt
with open('review_prompt.txt', 'r') as f:
prompt = f.read()

client = OpenAI(
base_url="https://openrouter.ai/api/v1",
api_key=api_key
)

try:
response = client.chat.completions.create(
model=model,
messages=[
{
"role": "user",
"content": prompt
}
],
extra_headers={
"HTTP-Referer": "https://github.com",
"X-Title": "AI PR Review"
}
)

# Save the model used to a file
with open('review_model.txt', 'w') as f:
f.write(response.model)

print(response.choices[0].message.content)
return 0
except Exception as e:
print("## ⚠️ AI Review Failed")
print(f"Error: {str(e)}")
print("")
print("This could be due to:")
print("- API rate limiting")
print("- Large diff size")
print("- Temporary service issues")
print("")
print("Please retry the review later or request manual review.")
return 1

if __name__ == "__main__":
sys.exit(main())
EOF

# Run AI review with timeout
TIMEOUT_SECONDS=$((${AI_EXECUTION_TIMEOUT_MINUTES} * 60))
echo "🤖 Using $AI_MODEL via OpenRouter for PR review (timeout: ${AI_EXECUTION_TIMEOUT_MINUTES} minutes)..." # pragma: allowlist secret
if ! timeout $TIMEOUT_SECONDS python ai_review.py > review_output.md; then
echo "## ⚠️ AI Review Failed" > review_output.md
echo "AI review could not be completed (timeout or error)." >> review_output.md
echo "" >> review_output.md
echo "This could be due to:" >> review_output.md
echo "- API rate limiting" >> review_output.md
echo "- Large diff size" >> review_output.md
echo "- Temporary service issues" >> review_output.md
echo "" >> review_output.md
echo "Please retry the review later or request manual review." >> review_output.md
echo "ai_review_failed=true" >> $GITHUB_OUTPUT
else
echo "ai_review_failed=false" >> $GITHUB_OUTPUT
fi

- name: Post review comment
if: (steps.check_tests.outputs.result == 'true' || github.event_name != 'pull_request') && steps.check_recent_review.outputs.skip != 'true'
uses: actions/github-script@v7
AI_ENABLE_INLINE_COMMENTS: ${{ vars.AI_ENABLE_INLINE_COMMENTS || 'true' }}
with:
github-token: ${{ github.token }}
script: |
const fs = require('fs');
const review = fs.readFileSync('review_output.md', 'utf8');
let model = 'unknown';
try {
model = fs.readFileSync('review_model.txt', 'utf8').trim();
} catch (e) {
console.log("Could not read model file, using 'unknown'.");
}
github_token: ${{ github.token }}
openrouter_api_key: ${{ secrets.OPENROUTER_API_KEY }}
model: ${{ vars.AI_MODEL || 'anthropic/claude-sonnet-4' }}
review_type: 'full'
max_tokens: 4096
temperature: 0.7
request_timeout_seconds: 600 # 10 minutes default, can be overridden by the action
retries: 3
post_comment: 'true' # Let the action handle resolvable comments

const comment = `## 🤖 AI Review\n\n${review}\n\n---\n` +
`*This review was automatically generated by \`${model}\` via OpenRouter. Please consider it as supplementary feedback alongside human review.*`;

// For PR events
if (context.eventName === 'pull_request') {
await github.rest.issues.createComment({
...context.repo,
issue_number: context.issue.number,
body: comment
});
}
// For issue comments on PRs (when /review is used)
else if (context.eventName === 'issue_comment') {
await github.rest.issues.createComment({
...context.repo,
issue_number: context.issue.number,
body: comment
});
}

- name: Handle AI review failure
if: steps.ai_review.outputs.ai_review_failed == 'true'
if: steps.ai_review.outputs.review_status == 'failure' || steps.ai_review.outputs.review_status == 'error'
uses: actions/github-script@v7
with:
github-token: ${{ github.token }}
Expand All @@ -386,6 +203,13 @@ jobs:
labels: ['ai-review-failed']
});

// Post a comment explaining the failure
await github.rest.issues.createComment({
...context.repo,
issue_number: context.issue.number,
body: `## ⚠️ AI Review Failed\n\nThe AI review could not be completed. Status: ${{ steps.ai_review.outputs.review_status }}\n\nThis could be due to:\n- API rate limiting\n- Large diff size\n- Temporary service issues\n\nPlease retry the review later or request manual review.`
});

// Fail the workflow step to indicate the review failure
core.setFailed('AI review failed - manual review needed');

Expand All @@ -394,13 +218,14 @@ jobs:
github.event_name == 'pull_request' &&
steps.check_tests.outputs.result == 'true' &&
steps.check_recent_review.outputs.skip != 'true' &&
steps.ai_review.outputs.ai_review_failed != 'true'
steps.ai_review.outputs.review_status == 'success'
uses: actions/github-script@v7
env:
REVIEW_COMMENT: ${{ steps.ai_review.outputs.review_comment }}
with:
github-token: ${{ github.token }}
script: |
const fs = require('fs');
const review = fs.readFileSync('review_output.md', 'utf8').toLowerCase();
const review = process.env.REVIEW_COMMENT.toLowerCase();

const labels = [];

Expand Down Expand Up @@ -458,7 +283,7 @@ jobs:
github.event.label.name == 'ai-review-needed' &&
steps.check_tests.outputs.result == 'true' &&
steps.check_recent_review.outputs.skip != 'true' &&
steps.ai_review.outputs.ai_review_failed != 'true'
steps.ai_review.outputs.review_status == 'success'
uses: actions/github-script@v7
with:
github-token: ${{ github.token }}
Expand Down
Loading
Loading