Browse by task

Start from the job you need done. Each list is curated, sourced and comes with a verdict.

• Issue → PR8Agents that take a well-written issue, work in your repo and hand back a pull request you review.
• Repo-aware coding12Coding agents that read the whole codebase, edit multiple files and run commands, not just autocomplete a line.
• Code review3AI reviewers and review workflows that catch bugs and risky changes before merge.
• Testing & QA5Ways to make agents check their own work: run tests after every edit, click through the UI in a real browser, and run security checks before launch.
• Agent context & rules8Rules files, conventions, skills and up-to-date docs that stop agents from guessing.
• Deep research5Agents that plan a research question, read many sources and write a cited report while you do something else.
• Web scraping & browsing7Tools that turn web pages into clean, LLM-ready data or let an agent operate a browser.
• Structured extraction5Getting clean text and structured fields out of web pages and documents so models can use them.
• Automation pipelines6Platforms and workflows that run recurring jobs with AI steps: triage, enrichment, digests and reports.
• Multi-agent handoff9Frameworks and patterns for agents that delegate to specialists, hand off work and keep state.
• MCP integrations10Model Context Protocol servers that connect agents to GitHub, databases, browsers, docs and automations.
• Prompt-to-deploy6Builders that turn a description into a running, deployed web app.
• Internal tools3Fast ways to build dashboards, bots and small apps for your own team, where speed matters more than polish.
• Local & BYO models5Open-source tools that let you choose your model provider, including local models.
• Production guardrails6Principles, sandboxes, hooks and observability that keep agents predictable once real users and real data are involved..