Skip to content
SkillModel-led1,724 tokens in use

roguelike

Designs run-based roguelikes and roguelites in any engine: run structure, permadeath, meta-progression, sub-seeded and daily runs, energy turn order and symmetric field of view.

9,683 stars on the repository Apache License 2.0gamedev-skills/awesome-gamedev-agent-skills · skills/genres/roguelike
npx skills add gamedev-skills/awesome-gamedev-agent-skills
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing roguelike costs

94tokens idle
1,724tokens in use
3,462tokens, up to

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Up to adds the files its instructions link. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

Nothing detected: no shell, network, file writes, git or credentials in its files.

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Excellent
  • Frontmatter8/15
  • Structure25/25
  • Examples9/15
  • Output10/15
  • Size15/15
  • Resources15/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent

No findings.

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Good
  • Development activity92
  • Maintenance96
  • Community60
  • Maturity24
  • Momentum40

Roguelike owns the run: what a run contains, what survives death, how turns and seeds work, and what a generated floor must guarantee. Dungeon generators come from the procedural-gen skill; this skill decides what they must produce and checks it.

The agent settles whether the game is a traditional roguelike or a roguelite, lays out acts, splits run state from profile state, seeds every floor from one run seed and schedules turns with an energy system. It then runs tests on seed determinism, floor acceptance, FOV symmetry, speed ratios and the death save rule.

What You Get

  • A run layout of acts or biomes with a reward choice every few minutes and a first death around 5 to 15 minutes in
  • A split between run state and profile state, with a written permadeath rule
  • Per-floor, per-purpose sub-seeds from one run seed
  • An energy scheduler pattern with discrete speeds of 50, 100 and 200
  • A test list: seed hashes, 500 or more seeds per depth, FOV symmetry, speed ratios and death save handling
  • A reference on floor acceptance checks, daily seeds, symmetric shadowcasting and spawn budgets

Common Use Cases

  • Designing a NetHack or Hades style run with a death screen and unlocks
  • Making the same seed give the same floors even after different play
  • Fixing monsters that never act or enemies that see the player unseen
  • Setting up a daily seed with a shared leaderboard

What Makes It Unique

It treats a run as a set of guarantees that can be tested. The same seed must give the same floor hash after different play on floor 1, hundreds of seeds per depth must pass acceptance checks, and dying must leave the profile file unchanged. It also names causes of subtle bugs, such as a language hash that changes per process.

Under The Hood

The skill is a SKILL.md with one reference file on seeds, FOV and acceptance checks. The agent reads your game code, edits it and writes tests, such as same-seed hash comparisons and speed ratio checks, then runs them with the project’s test runner. The snippets are pseudocode, and one gotcha names a python-tcod FOV call. The version notes cite outside pages, but the skill does not tell the agent to fetch web pages or call services.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers