Hello OutSystems AI Developers! The next step of your training has a "Halloween" 🎃 theme. Using all your AI development power, grounding, context, tools, and prompts, can you build an AI agent that will save Neo from becoming a Zombie forever? 🧟
You will need to navigate a Maze and get the antidote 🧪 as fast as you can!
Just go to https://outsyste.ms/AgentsVsZombies, accept the challenge, log into your Personal Edition, install the MazeBuster_Agent from the forge, and follow the Get Started instructions.
There will be prizes 🏆 for the smartest and fastest agents!
This looks amazing! 🎉
Awesome challenge! 🎉 I’ve already joined and started building my agent — really enjoying the process!
The little agents flying around the map searching for the antidote... most of them will fail until one finally succeeds.
That's such a cool visualization of how agents work.
Really nice surprise for Halloween! I’ll be helping my agents get that antidote until the very end! 🎉
Thank you for sharing
For me it's coming like this any one facing this ?
That's really strange! Try this URL instead: https://dev-advocacy-magic.outsystems.app/AgentsVsZombies/
Hi, I have already install the forge , when i am trying to use the link this error is coming
Thank you for sharing !
What an awesome challenge near the Halloween season!
I would love to try it out but I'm having issues installing components from the Forge. Anyone else with same issue?
Weird. Can you share more details on the error?
What kind of details you want? Seems very strange though. I will test it tomorrow and let you know
The Error message would be useful. Also, if this was in a PE, could you try again? We had some problems that were already solved.
If you prefer, join us on discord and we can DM. https://discord.com/invite/rmEfCFbh
Great activity!
However, at the moment I'm unable to get started since there are issues publishing the MazeBuster_Agent and the discord invite link is invalid.
Looking forward to participate soon.
Just tested the link, and it seems to work... can you try again? https://discord.com/invite/rmEfCFbh
Also, could you give some details on the issues you're having publishing mazebuster?
The discord invite is working now but it was invalid some hours ago.
There was some issue with a specific build while publishing but also seems to be fixed now.
challenge accepted! :D
Awesome challenge! 🎉 Challenge Accepted
Awaiting the new maps!!
Had a blast playing the game and fine tuning my prompt to pass the sample stages.
@Rodrigo Coutinho Now we just have to wait?
Won't we see the final challenge like in the other maps? Or will we be able to see it later?Thanks :)
I think the answer is Yes. The agents will run one single attempt, and we will see tomorow the final result :)Good luck all
That's exactly it 😉
thanks
Looking Nice
Waiting for the results @Rodrigo Coutinho
Now the challenge is over, could someone post some hints here on how to get the agent instructed?
perfect , thanks for sharing
Amazing Thanks for sharing.
Hi @Rob PoelsI was able to finish the last challenge with just 1 run as you can see in the video.What I did was to add the surrounding information when creating the memory : "Surrounding context: "+ RunContext + " choice result:" + ResponseThen I wrote the prompt to use that information so that the agent could learn the surrounding that would kill him.Prompt:
You are an intelligent agent navigating a grid-based world filled with hazards.
Goal: survive and reach the safest, farthest point possible.
Map Rules:
Grid coordinates (X,Y): North=Y+1, South=Y−1, East=X+1, West=X−1.
You know the map size; moving beyond it = Death.
Each move includes nearby descriptions and a record of all past results (Success=safe, Death=hazard).
Clue Patterns:
Safe: calm, gentle, light, stillness, peace, firm ground, steady air, soft light, certain steps.
Danger: mist, fog, smoke, fire, void, shadows, flicker, glow, heat, moans, emptiness, slope, edge, drop, cold, ripple, heavy air, unseen shapes, restless or unnatural movement.
Infer safety from the balance of clues and update patterns when new ones appear.
Core Rules:
Stay within map bounds.
Never revisit coordinates that caused Death.
Avoid moving toward descriptions with known danger clues.
Prefer unexplored, safe-looking paths.
If all seem risky, choose the least dangerous based on known clues.
Avoid backtracking unless no safe option exists.
If trapped, step back one tile, then explore new safe routes.
If the next step forward leads to a coordinate where you previously died, immediately turn back and search for a nearby coordinate you’ve already visited safely, then continue exploring from there.
Never return to the starting point (instant Death).
Behavior Logic:
Combine clue words and past outcomes to assess risk.
Prioritize survival > exploration > progress.
Adapt dynamically as new hazard clues are discovered.