diff --git a/guides/20260531_openhands_pro_tips.md b/guides/20260531_openhands_pro_tips.md
new file mode 100644
index 00000000..c853976e
--- /dev/null
+++ b/guides/20260531_openhands_pro_tips.md
@@ -0,0 +1,176 @@
+# OpenHands Pro Tips: Maximizing Your AI Development Efficiency
+
+OpenHands is powerful out of the box, but the real speed comes from using it like a production tool instead of a novelty demo. If you treat it like a serious development environment—tight task scope, good prompts, clean project setup, and fast feedback loops—you can cut a lot of wasted time.
+
+This guide collects practical tips, hidden shortcuts, and workflow habits that help data teams, ML engineers, and tech leads get more out of OpenHands.
+
+## Why efficiency matters
+
+AI development environments are only useful if they reduce context switching. The goal is not to make the agent do everything blindly. The goal is to make it easier to:
+
+- start with the right repo and branch
+- keep tasks small and verifiable
+- avoid repeating the same setup work
+- catch mistakes early
+- ship changes with less friction
+
+The more structure you give OpenHands, the better the output.
+
+## 1. Start with a narrow task
+
+The most common mistake is asking for too much at once.
+
+Bad:
+
+> Fix the whole app, improve performance, and add tests.
+
+Better:
+
+> Inspect the login flow, identify the slowest function, and add a test for the regression.
+
+Small tasks make it easier for OpenHands to reason about the repo and give you a clean result. You can always chain the next task after the first one is verified.
+
+## 2. Give the agent concrete acceptance criteria
+
+A good prompt includes a clear end state. That saves a lot of back-and-forth.
+
+Example:
+
+> Update the data loader so it caches responses for 10 minutes. Add a unit test that verifies cache expiry. Do not change unrelated code.
+
+That makes it easier to tell whether the task is done and reduces accidental scope creep.
+
+## 3. Use repo-specific setup once, then reuse it
+
+If your team repeatedly works in the same repo, standardize the startup steps:
+
+- default branch
+- install commands
+- environment variables
+- test command
+- lint command
+- common folders to avoid
+
+You can keep this as a reusable checklist or project note so each session starts faster.
+
+## 4. Keep your workspace clean
+
+A cluttered repo slows the agent down.
+
+Best practices:
+
+- remove dead branches and stale temp files
+- keep `.env` and secrets out of the workspace
+- use a consistent folder layout
+- avoid committing generated artifacts unless needed
+- isolate feature work in a branch per task
+
+A clean workspace means fewer false positives and less time spent looking in the wrong place.
+
+## 5. Prefer fast feedback loops
+
+The most efficient OpenHands sessions follow this loop:
+
+1. inspect the issue
+2. make the smallest useful change
+3. run tests or validation
+4. fix failures immediately
+5. repeat until green
+
+Don’t wait until the end to test everything. Catching issues in the first 5 minutes saves a lot of rework.
+
+## 6. Use task decomposition for larger work
+
+For bigger work, split the job into layers:
+
+- discovery: understand the repo and issue
+- implementation: make the code change
+- validation: run tests and fix failures
+- packaging: write the summary or PR description
+
+That structure helps the agent stay focused and makes it easier for humans to review the result.
+
+## 7. Avoid prompt bloat
+
+Long, vague prompts are slower and often worse.
+
+Instead of pasting your life story, include only what matters:
+
+- file paths
+- error messages
+- relevant constraints
+- the desired outcome
+
+The cleaner the prompt, the faster the agent can act.
+
+## 8. Use the right level of automation
+
+Not every step should be automated the same way.
+
+Good candidates for automation:
+
+- code formatting
+- test runs
+- repetitive refactors
+- file moves
+- dependency updates
+
+Better left to a human review loop:
+
+- architectural changes
+- security-sensitive edits
+- deployment decisions
+- large behavior changes
+
+The fastest teams mix automation with deliberate checkpoints.
+
+## 9. Track the common failure modes
+
+OpenHands users tend to hit the same mistakes repeatedly:
+
+- running in the wrong branch
+- forgetting to install dependencies
+- changing unrelated files
+- skipping tests
+- over-editing instead of fixing the root cause
+
+If you notice one of these patterns, write it down and turn it into a team rule.
+
+## 10. Build reusable conventions
+
+The biggest productivity gains come from conventions that reduce decision-making:
+
+- one branch per issue
+- one test command per repo
+- one lint command per repo
+- one review checklist per PR
+- one place for environment notes
+
+Once the team agrees on those defaults, the agent can move faster with fewer interruptions.
+
+## Practical workflow example
+
+Here’s a simple workflow for a new issue:
+
+1. Open the repo and the issue.
+2. Identify the exact file or function to change.
+3. Make the smallest safe fix.
+4. Run the relevant test target.
+5. If the test fails, inspect the failure and patch it.
+6. Write a clear PR summary explaining what changed and why.
+
+That basic loop is boring—but it works.
+
+## Common pitfalls to avoid
+
+- Don’t ask OpenHands to solve a giant multi-part problem in one pass.
+- Don’t hide important context in a huge wall of text.
+- Don’t skip verification because the code “looks right.”
+- Don’t change unrelated files unless there’s a strong reason.
+- Don’t treat the first answer as final if tests disagree.
+
+## Final takeaway
+
+OpenHands is most effective when you use it like a disciplined engineering assistant. Give it a clear goal, keep the task small, validate quickly, and reuse your repo conventions. That’s how you turn it from a cool demo into a real productivity tool.
+
+If your team runs AI development work regularly, the real win is not just writing code faster—it’s reducing the amount of time spent figuring out what to do next.