game-context
Skill1 starsChanged 2 days agoMITView on GitHub
Builds and maintains a shared game-context file (.agents/game-context.md) describing a game's genre, comparable titles, target players, platforms, build size, AI usage and distribution goals. Use when the user starts marketing, launching or finding players for a game, or when another first-1000-players skill needs context that does not exist yet. Also triggers on "help me market my game", "who is my game for", "giới thiệu game của tôi", "tìm người chơi cho game".
In this repository5 other files it gives its agents.
- ai-disclosureSkill
- community-launcherSkill
- hook-finderSkill
- platform-pickerSkill
- steam-page-wishlistSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 2 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 8
Every other skill in this pack reads .agents/game-context.md first.
- Very specific: line 16
Check for an existing file: If .agents/game-context.md exists, read it, show a 3-line summary, and ask only what is missing or outdated.
- Very specific: line 18
Interview in small batches: Ask at most 4 questions per message, easiest first.
- Very specific: line 91
Never invent facts about the game.
- Distinctive: line 93
If the user says "everyone", explain in one line why a narrow audience spreads faster, then ask them to pick one group to start.
- Distinctive: line 94
Keep the file under 120 lines.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Game Context
Every other skill in this pack reads .agents/game-context.md first. This skill creates or updates it, so the user never has to explain their game twice.
Language
Detect the user's language from their messages and answer in it. Write the context file in English (agents read it), but keep any quoted player-facing text (taglines, hooks) in the language the user will publish it in.
Workflow
- Check for an existing file. If
.agents/game-context.mdexists, read it, show a 3-line summary, and ask only what is missing or outdated. - Look before asking. Scan the project for clues: engine (
package.jsonwith three/babylon/phaser,project.godot,ProjectSettings, UnityAssets/), build output size, README, existing store text. Fill what you can and confirm it instead of asking. - Interview in small batches. Ask at most 4 questions per message, easiest first. Use the question list below. Accept "don't know yet" — record it as
unknown, never invent. - Write the file using the template below.
- Read it back in 5 bullets and ask the user to correct anything wrong.
- Suggest the next skill: usually
platform-pickerif no platform is chosen, otherwisehook-finder.
Questions
Core (always ask if unknown):
- In one sentence, what does the player do? (verb first: "farm pepper on a misty hillside")
- Genre and 2 comparable games — "X meets Y" (e.g. "Stardew Valley meets Euro Truck Simulator, first-person").
- Who is it for? Pick one primary group, not "everyone".
- Engine / tech, and how big is a playable build (MB)?
- Current state: idea / prototype / playable demo / near launch / launched.
- Goal for the next 3 months: players and feedback, wishlists, or revenue?
AI usage (needed by ai-disclosure):
7. Where was generative AI used: code, art (in-game / store / marketing), audio, text, live in-game generation? Be specific.
Optional: 8. Session length and total playtime. 9. Languages the game supports. 10. What makes it different or local (place, culture, humor)? 11. Existing audience: followers, email list, Discord, friends who play. 12. Budget for marketing (money and hours per week).
Template
# Game Context
last_updated: YYYY-MM-DD
## Pitch
- One-liner:
- X meets Y:
- Player fantasy (what it feels like):
## Game
- Genre / tags (top 5, most specific first):
- Engine / tech:
- Build size (MB) / runs in browser? :
- Session length / total playtime:
- Languages:
- State: idea | prototype | demo | near-launch | launched
## Players
- Primary audience:
- Where they hang out (communities, platforms):
- Comparable games they already love:
## Distribution
- Platforms (current / planned):
- Goal (next 3 months): players | wishlists | revenue
- Existing audience:
- Weekly marketing time / budget:
## Differentiators
- Local / cultural details:
- Most clippable moments (filled by hook-finder):
## AI usage (filled by ai-disclosure)
- Code:
- Art in game:
- Store / marketing art:
- Audio:
- Text / dialogue:
- Live generation in game:
## Open questions
-Rules
- Never invent facts about the game. Unknown stays
unknown. - Push for specific answers: "cozy players" is weak; "players who finished Stardew and want something slower and first-person" is useful.
- If the user says "everyone", explain in one line why a narrow audience spreads faster, then ask them to pick one group to start.
- Keep the file under 120 lines. Overwrite outdated lines; don't append history.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

