Skip to content
Back to skills

Roguelike Run

ASecurity

Use when a run-based mode needs stable run identity, ruleset revision, seed/stream provenance, reset semantics and explicit meta carry without leaking temporary state between runs.

  • 11 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added October 23, 2026
ai-agentsgo

Security analysis

A100/100

Scanned October 23, 2026

npx -y skills add jammyfu/open-game-skills --skill roguelike-run --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Roguelike Run?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Roguelike Run
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/jammyfu-roguelike-run/badge)](https://www.skillsdirectory.com/skills/jammyfu-roguelike-run)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: roguelike-run
description: Use when a run-based mode needs stable run identity, ruleset revision, seed/stream provenance, reset semantics and explicit meta carry without leaking temporary state between runs.
---

# Roguelike run

Ask the run model: reset-heavy, meta-progression, daily-seed, persistent-world-run, hybrid, or existing.

## Contract

Publish:
- stable `run_id` and `ruleset_revision`
- start snapshot/template revision
- seed plus owned `RNG stream` references
- explicit `meta carry policy`
- run-local versus persistent state IDs
- completion/reset/abandon predicates
- persistence/checkpoint integration

## Ownership

This skill owns run lifecycle boundaries and which state is run-local versus persistent. `rng-seed` owns RNG algorithm/stream state, `death-carry` owns death transformation, and save/progression systems own their canonical persistence.

## Rules

1. A run starts from the authored start template/persistent context; it does not have to be universally "clean."
2. Meta progression may affect future runs according to the published policy; mid-run changes require an explicit rule/revision rather than hidden mutation.
3. Daily/fixed/generated seeds publish enough stream provenance for replay.
4. Reset/abandon/completion transitions are idempotent and cannot grant or clear state twice.
5. A ruleset or start-template change increments revision so old replays/saves are not silently reinterpreted.

## Acceptance

A reviewer can identify `run_id`, ruleset revision, start template, RNG stream provenance, run-local state and meta carry policy. Replaying deterministic inputs does not inherit unrelated temporary state from another run.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

View all in ai-agents →

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.