By Mike Yacullo

I first came across Action / Theme Oracles in Shawn Tomkin’s fantastic, free solo RPG Ironsworn. He used them again in his equally excellent but not free sci-fi follow-up, Starforged. These are tables that you roll on that will give you a verb (Action) and a noun (Theme) that you then combine and interpret within the context of your situation. They work well to answer questions like “What’s this NPC’s motivation?” or “What should be the general theme of my next adventure?”

Let’s use the NPC motivation example. You might roll on the tables and get “Oppose” on the Action table, and “Superstition” on the Theme table. So you think about what “Oppose Superstition” might mean during your adventure. Maybe the character is trying to bring science to a less-developed civilization. Maybe they are trying to reveal that allegedly supernatural occurrences have a more mundane cause. It all depends on context.

In my game, my character came across a research station studying a crashed alien probe. What was the purpose of the probe? I went to the Action / Theme table and my rolls resulted in “Serve Labor” – I interpreted this to mean that the probe has enslaved the scientists to force them to build something.

These tables can work right out of the box with Captain’s Log – especially the ones in Starforged, which I encourage everyone to buy. But I wanted some tables that were more specifically Trek-oriented. I was sure GPT-4 could help me out here.

[Note: as in my previous article, I am using a paid $20/month GPT-4 subscription, but you could take the same approach with GPT-3.5 or other free AIs]

As usual, I began by asking GPT-4 if it was familiar with the subject of our discussion. Throughout the document, GPT-4’s responses are in typewriter font.

Great, it has a good grasp of the concept, so we can move forward.

  1. Negotiate – Engage in diplomatic discussions or peace talks.
  2. Explore – Discover a new planet, star system, or anomaly.
  3. Defend – Protect your ship or an ally from an attack.
  4. Repair – Fix a critical system failure or damage to the ship.
  5. Investigate – Solve a mystery or gather information.
  6. Rescue – Save individuals from danger or captivity.
  7. Patrol – Monitor a sector for unusual activity.
  8. Research – Conduct scientific experiments or studies.
  9. Pursue – Follow a suspect or enemy.
  10. Evade – Escape from a pursuit or dangerous situation.
  11. Mediate – Resolve a conflict between factions or species.
  12. Trade – Engage in commerce or negotiation for resources.
  1. Diplomacy – Interactions focus on peacekeeping and alliances.
  2. Mystery – Something unknown or puzzling is central.
  3. Conflict – A struggle or battle, possibly involving phasers or starships.
  4. Discovery – Finding something new or unexpected.
  5. Survival – Overcoming a life-threatening situation.
  6. Honor – A matter of ethical, moral, or cultural significance.
  7. Betrayal – Trust is broken, or a deception is revealed.
  8. Science – The pursuit of knowledge is key.
  9. Alien Culture – Interaction with or understanding of a foreign species.
  10. Espionage – Covert operations or intelligence gathering.
  11. Time Anomaly – Events involving time travel or temporal distortions.
  12. Anomaly – A strange or unusual phenomenon that defies explanation.

A very good start! Now to fine-tune the results. There were a couple of issues, but luckily you can just talk to GPT-4 and tell it what you want changed:

The next set of results (not shown here) were very close to what I wanted, but there were a few anomalies:

  • The list went along for a while and then partway through GPT-4 seemed to decide that it should be in alphabetical order. Even though I know this doesn’t actually make a difference, I want the choices to be in a random, non-alphabetical order
  • There were a few odd choices (such as Zen and Zephyr) that snuck in. I suspect this is somehow tied to the alphabetical order. 

The final change I made was to reduce the number of items to 20. I created mine with 100 items to match Ironsworn and rolling d100, but I felt like the quality began to drop.

With this, I was happy with the results. I now had an Action / Theme matrix I could use to spark (not replace!) creativity in my Captain’s Log game. 

Here is a condensed prompt, combining all my trial and error into one, and the results:

  1. Negotiate
  2. Explore
  3. Beam
  4. Scan
  5. Infiltrate
  6. Repair
  7. Analyze
  8. Command
  9. Evade
  10. Rescue
  11. Diplomacy
  12. Adapt
  13. Decipher
  14. Innovate
  15. Sabotage
  16. Navigate
  17. Construct
  18. Detect
  19. Communicate
  20. Strategize

Theme Oracle (Star Trek Themed)

  1. Discovery
  2. Anomaly
  3. Federation
  4. Galaxy
  5. Quantum
  6. Nebula
  7. Odyssey
  8. Artifact
  9. Paradox
  10. Alien
  11. Diplomacy
  12. Unity
  13. Sovereignty
  14. Technology
  15. Intrigue
  16. Betrayal
  17. Peace
  18. Enlightenment
  19. Frontier
  20. Science

AI is unpredictable, so you may need to run this multiple times before you get the results you like. Some ways you may consider tweaking the prompt:

  • Ask for summaries of each word
  • Ask for different elements to be included or excluded

The nice thing is, because these are so easy to generate when you feel like you’ve seen most of the “good” combinations, you can create a new one. You can even create oracles for specific situations:

  1. Ambush
  2. Reinforce
  3. Infiltrate
  4. Sabotage
  5. Negotiate
  6. Defend
  7. Intercept
  8. Retreat
  9. Spy
  10. Strategize
  11. Alliance
  12. Rescue
  13. Patrol
  14. Encrypt
  15. Outmaneuver
  16. Assault
  17. Evacuate
  18. Blockade
  19. Collaborate
  20. Propagandize
  1. Dominion
  2. Federation
  3. Cardassia
  4. Klingon
  5. Romulan
  6. Bajor
  7. Betrayal
  8. Alliance
  9. Espionage
  10. Occupation
  11. Rebellion
  12. Truce
  13. Starfleet
  14. Maquis
  15. Section31
  16. Ferengi
  17. NeutralZone
  18. Warfront
  19. Resistance
  20. Treaty

In this way, you can be specific or broad, and get ideas no matter what your era and setting.

Just for fun, I wanted to see what GPT-4 would make of the results from the Dominion War matrix:

As with any matrix, if you roll a combination that doesn’t make sense or sound fun to you, re-roll, or pick the combo that sparks your imagination.

I hope this article has helped you see new ways to move your game forward. I’m happy to answer questions in the comments.

Good luck, Captains!


TM & © 2026 CBS Studios Inc. © 2026 Paramount Pictures Corp. STAR TREK and related marks and logos are trademarks of CBS Studios Inc. All Rights Reserved.

2 responses to “Using ChatGPT to Generate Action / Theme Matrixes and Scenarios for Captain’s Log”

  1. ChatGPT also understands the concept of a probability matrix, I use it occasionally to create random events for my tundra planet colony solo campaign.

    1. Michael Dismuke Avatar
      Michael Dismuke

      That is so cool!

Leave a Reply to Rudolf StolteCancel reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Discover more from Continuing Mission

Subscribe now to keep reading and get access to the full archive.

Continue reading