Andrey Markin
  • home
  • blog
  • projects
  • services
  • directory
    • Tool
    • Library
    • Repo
    • Article
    • XTweet
    • Guideline
  • about
  • contact
  • meet

Mark Life Ltd

  1. Home
  2. Directory
  3. Pullfrog
Meet

Mark Life Ltd

BG208147965

HomeContactPrivacyLLM-friendlyBlog RSSDirectory RSS
ToolAI CodingDev Toolsagentsci

Pullfrog

AI PR review agent from Zod's creator. Dynamically scales effort, uses subagent swarms for large PRs, and runs inside GitHub Actions with batteries-included MCP tools.

Added May 15, 2026Published May 12, 2026
Visit tool

PR review agent built by Colin McDonnell (creator of Zod). Runs inside GitHub Actions, dynamically scales effort to PR size, and goes beyond review with arbitrary tasks via @pullfrog mentions.

Code review

  • Dynamically scales effort
  • Sub-30s reviews for small PRs
  • Subagent swarms for large PRs
  • Incrementally reviews new commits
  • Maintains PR context across runs
  • One-click Fix button on actionable reviews
  • Auto-resolves addressed comments

Beyond PR review

Tag @pullfrog in GitHub with arbitrary tasks or configure automations:

  • Review new PRs
  • Address human review
  • Triage, reproduce, and label new issues
  • Auto-fix CI failures and merge conflicts
  • Custom cron tasks

Batteries included

The agent harness ships with built-in MCP tools and skills:

  • Git and GitHub operations
  • Headless browser and screenshots
  • Dependency auto-installation
  • Repo-level learning and memory
  • S3 and file upload

Lifecycle hooks

Runs inside Actions and can invoke your repo's linter, test suite, and other scripts via setup, post-checkout, pre-push, and post-run hooks.

Security

  • Isolated shell tool — no secrets or git creds exposed to the agent
  • GitHub tokens are run-scoped and auto-revoked
  • Built-in branch protections, no pushes to main
  • Secret auto-masking

Billing

First 30 runs free, then 7¢/run plus model costs. Use Pullfrog Router (any model at cost, no markup), bring your own key, or a Claude Code subscription.

Related

  • Using Claude Code: The Unreasonable Effectiveness of HTMLThariq makes the case for HTML over Markdown as Claude Code's output format — richer visualizations, color, diagrams, and easier sharing once agents are the ones authoring and humans rarely hand-edit.
  • quality-codeRhys Sullivan's TypeScript coding skill — branded types, discriminated unions, end-to-end types, real tests over mocks, OpenTelemetry observability, and picking the right abstractions instead of premature ones.
  • ExecutorOpen-source tool gateway for agents from Rhys Sullivan. Normalizes MCP, OpenAPI, GraphQL, and custom sources into one typed SDK that agents call in a sandbox — built on Cloudflare's Code Mode pattern.
  • improve-codebase-architectureMatt Pocock's skill for deepening modules — restructuring prototype code into something long-lived with cleaner layer dependencies and easier navigation.