GitHub Assistant ​

See More

Install the GitHub Assistant to get:

  • PR reviews: multi-dimensional reviews posted on every pull request
  • Issue auto-responses: code-grounded answers and triage on new issues
  • Auto-labeling: consistent tags for PRs and issues based on code analysis
  • Conversational follow-ups: ask questions in PR/issue threads and get contextual replies

Everything runs inside your GitHub Actions runner, so responses are grounded in your repository. You're in full control: keep the defaults, add new rules, or disable checks you don't want. No required third-party services beyond your chosen LLM provider, and you can use any LLM supported by Visor.

Quick Start ​

Add this GitHub Action to your repository:

yaml
name: Visor
on:
  issues:
    types: [opened]
  pull_request:
    types: [opened, synchronize]
  issue_comment:
    types: [created]

jobs:
  visor:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: probelabs/visor-action@v1
        with:
          github-token: ${{ secrets.GITHUB_TOKEN }}

That's it. Visor will now automatically review PRs and respond to issues.

Defaults and Overrides ​

The GitHub Assistant uses Visor's default workflow configuration: defaults/visor.yaml

To extend or customize code review steps, see Intelligent Code Review.

Issue Auto-Response ​

When users open issues, Visor analyzes your codebase and provides helpful responses grounded in actual code:

Issue #234: How do I configure rate limiting?question
U
@user opened this issue 2 minutes ago

I'm trying to set up rate limiting for my API endpoints but can't find the configuration options.

V
visor-bot commented just now

Rate limiting is configured in config/api.yaml:

rate_limit:
  enabled: true
  requests_per_minute: 100

See src/middleware/rate_limiter.go:45 for implementation details.

Related: #189, #201

PR Reviews ​

Multi-dimensional code review covering security, performance, architecture, and code quality - automatically posted to your PRs:

PR #456: Add user authentication+342 -28
Vvisor-bot reviewed

Security Review

  • JWT validation properly implemented
  • Password hashing uses bcrypt with appropriate cost
  • No hardcoded secrets detected

Performance Review

  • Consider caching user lookups (src/auth/user.go:78)
Suggested change
src/auth/middleware.go:45
- if err != nil { return }
+ if err != nil { return fmt.Errorf("auth: %w", err) }

For detailed review configuration, see Intelligent Code Review.

Auto-Labels ​

Automatically categorize PRs and issues based on code analysis:

PR #457: Fix database connection pooling
Vvisor-bot added labels:bugdatabaseperformance

Auto-detected:

  • Modifies: src/db/pool.go, src/db/connection.go
  • Impact: High (core infrastructure)
  • Reviewers suggested: @db-team

Conversational Follow-ups ​

Mention @visor-bot in any comment to ask questions or get clarification:

Conversation in PR #456
D
@developer

@visor-bot can you explain why you suggested caching for user lookups?

V
visor-bot replied

The GetUser() function is called 3 times per request:

  1. middleware.go:45 - auth check
  2. handler.go:78 - permission check
  3. response.go:23 - user data serialization

Each call queries the database. A request-scoped cache would reduce DB calls from 3 to 1.

Example: src/cache/request_cache.go

GitHub Action Options ​

yaml
- uses: probelabs/visor-action@v1
  with:
    # Required
    github-token: ${{ secrets.GITHUB_TOKEN }}

    # Optional - GitHub App auth (better rate limits/permissions)
    app-id: ${{ secrets.PROBE_APP_ID }}
    private-key: ${{ secrets.PROBE_APP_PRIVATE_KEY }}
    installation-id: ${{ secrets.PROBE_APP_INSTALLATION_ID }}

    # Optional - custom config file (default: visor.yaml)
    config-file: .visor.yaml

    # Optional - AI provider (default: anthropic)
    ai-provider: anthropic

    # Optional - fail on critical/error issues (default: true)
    fail-on-issues: true

    # Optional - max parallel steps (default: 4)
    max-parallelism: 4

LLM Provider Configuration ​

bash
# Anthropic (Claude)
export ANTHROPIC_API_KEY=your-api-key

# Google (Gemini)
export GOOGLE_API_KEY=your-api-key

# OpenAI (GPT-4)
export OPENAI_API_KEY=your-api-key

# Azure OpenAI
export AZURE_OPENAI_API_KEY=your-api-key
export AZURE_OPENAI_ENDPOINT=https://your-resource.openai.azure.com

# AWS Bedrock
export AWS_ACCESS_KEY_ID=your-access-key
export AWS_SECRET_ACCESS_KEY=your-secret-key
export AWS_REGION=us-east-1

# Self-hosted (OpenAI-compatible)
export OPENAI_API_KEY=your-api-key
export OPENAI_BASE_URL=http://localhost:8080/v1

# GitHub Actions example (pick one provider)
# env:
#   ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}

Next Steps ​