Writing code, filing issues, opening PRs, reviewing changes — the repetitive tasks developers do every day can now be handed off to Openclaw with a single sentence.
These are real scenarios. You do not need to know git commands or look up docs. Just describe what you want, and Openclaw will handle the rest.
You told Openclaw
"Fix the null pointer crash on the login page in the main branch. Once fixed, open a PR with a clear title describing what changed."
What Openclaw did
Openclaw read the file, located the offending line, created a new branch, committed the fix, and opened a Pull Request on GitHub with a well-written description explaining what was broken and what the fix does.
You told Openclaw
"Based on Issue #42, add a delete account feature to the user settings page. Include tests and open a PR when done."
What Openclaw did
It read the issue, understood the requirements, wrote the component code, backend endpoint, and corresponding tests, then opened a PR with a description and test instructions.
You told Openclaw
"Review the 5 new PRs from today. Focus on security issues and obvious logic bugs. Give me a summary ranked by severity."
What Openclaw did
Openclaw pulled the diff for each PR, analyzed the code changes, flagged 2 potential authorization issues and 1 missing null check, then organized the findings into a prioritized review report.
You told Openclaw
"Take the product brief my PM sent and break it into 3 GitHub Issues. Add appropriate labels and milestone to each."
What Openclaw did
It split the vague brief into separate issues for frontend UI, API changes, and database migration, each with detailed acceptance criteria, correct labels, and milestone assignments.
You told Openclaw
"Read all PRs merged between v1.3.0 and v1.4.0 and write a public-facing changelog grouped by features, fixes, and breaking changes."
What Openclaw did
Openclaw queried GitHub for all merged PRs in that range, synthesized the commits and PR descriptions, and produced a clean release note ready to paste into the GitHub Release page.
You told Openclaw
"Go through the issues from the past week. Classify them into bug, feature, question, and docs. Flag any duplicates you spot."
What Openclaw did
It read 38 issue titles and bodies, grouped them by type, applied the correct labels, identified 4 duplicates, and suggested two of them be merged into a single tracking issue.
The GitHub Skill gives Openclaw a complete toolkit that covers most of what developers do on GitHub every day.
Read, create, and update files. Browse directory structure, search code, view commit history and file diffs.
Create branches, commit code, open PRs, post review comments, merge pull requests, and list all open PRs queued for review.
File new issues, add comments, apply labels, assign milestones, bulk close, and even generate code directly from an issue body.
Search across the entire repo for function names, patterns, or comments. Read PR diffs and understand code context to give meaningful suggestions.
Generate changelogs from PR history, create git tags, and publish release notes to the GitHub Release page without touching the UI.
Not just data retrieval. It actually reads your code, understands requirements, and executes autonomously — freeing you from routine developer busywork.
Getting Openclaw to truly "take over" GitHub operations requires five core modules working together.
The GitHub Skill is the entry point for the entire setup. Installed through ClawHub, it gives Openclaw a complete set of GitHub tools — telling the AI what operations it can perform and how to call each one. The Skill does not run on its own; it depends on underlying tools like the gh CLI to translate intent into real GitHub actions.
GitHub's official command-line tool gh is the execution layer that talks directly to the GitHub API. Openclaw calls gh commands via its exec tool to create PRs, modify files, view diffs, and carry out the full range of real operations. This is not a simulation — it works exactly like a real developer running commands in a terminal.
Openclaw never uses your GitHub password directly. Instead, it goes through an OAuth flow to obtain a scoped access token. You control exactly which repos it can access and what actions it is allowed to take. The GitHub Skill documentation explicitly notes that this permission model was designed so you can grant access without giving the agent full account control.
Opening a PR is not one action — it involves reading files, understanding logic, finding the issue, modifying code, creating a branch, committing, and writing a description. A standard AI chat ends after one step. Openclaw's planning layer chains all of these steps together, executes them autonomously, and retries or adjusts strategy when something fails.
For tasks like "read 50 PRs and generate a report," a single chat context is not enough. Openclaw's orchestration layer lets it keep working — reading data in batches, accumulating intermediate results, and merging them into the final output — without you watching every step. This transforms the AI from a one-shot assistant into a teammate that can sustain ongoing work.
Once connected, these prompts can be sent directly to Openclaw. Just swap in your repo name and issue number.
In [repo-name], find the bug in [file-path] related to [problem description] on the main branch. Fix it in a new branch and open a PR explaining what you changed and why.
Read Issue #[number] and implement the feature in [repo-name]. Once done, open a PR and include instructions on how to test it.
Review PR #[number]. Focus on security gaps, edge-case handling, and whether the code follows the style conventions in this project. Give me an ordered list of findings.
Read all merged PRs in [repo-name] between [v1.x.x] and [v1.y.y]. Generate a changelog grouped by new features, bug fixes, and breaking changes.
Explore More Cases
Give your AI assistant a soul by configuring its personality, context, and operational rules.
Use Gmail API and OAuth to let Openclaw read emails, prioritize them, and draft replies.
Use Bird Skill with your browser session to turn Openclaw X into one workflow for posting, replying, monitoring mentions, reading threads, and AI-driven X operations.
Connect your GPT, Gemini, and Claude plans to Openclaw so Openclaw Codex, Openclaw Gemini, and Openclaw Claude work without extra API token billing.
Learn how to turn repeated work into reusable skills so AI can execute with stable steps, rules, and guardrails.
Learn the two Openclaw Browser starting routes: Playwright MCP and browser-use, and how each one should be set up.
Learn how to install and set up the Openclaw automation system, from environment checks to essential CLI commands.
See how OpenClaw searched inventory, contacted dealerships, and negotiated across browser, email, and messaging channels to save $4,200.
See how Openclaw searches job boards across the web, tailors resumes and cover letters, sends bulk personalized applications, and triages inbox replies — automating the most repetitive parts of job hunting.
See how Openclaw searches listings across platforms, compares prices side by side, sends bulk inquiry messages, and negotiates rent with landlords — automating the most tedious parts of apartment hunting.
See how Openclaw securely connects to cloud servers, analyzes error logs, updates config, and recovers failed deploys to keep production healthy around the clock.
See how Openclaw coordinates health data, smart home devices, meal planning, and daily schedules to run your personal systems as one workflow.