Skip to content
Back to skills

Interruptible Strategy Controller

ASecurity

Select and interrupt prioritized symbolic NetHack strategies with explicit action queues.

  • 247 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added September 9, 2026
testingpython

Security analysis

A100/100

Scanned September 9, 2026

npx -y skills add VectorSpaceLab/AREX-Skill --skill interruptible_strategy_controller --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Interruptible Strategy Controller?

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

Security grade badge for Interruptible Strategy Controller
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/vectorspacelab-interruptible-strategy-controller/badge)](https://www.skillsdirectory.com/skills/vectorspacelab-interruptible-strategy-controller)

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

SKILL.md
---
name: interruptible_strategy_controller
description: Select and interrupt prioritized symbolic NetHack strategies with explicit action queues.
---
# Interruptible Strategy Controller

Use this skill when a symbolic NetHack agent must choose among non-atomic strategies such as exploration, combat, healing, or survival. Do not use it for combat scoring details; pass combat states to a scorer skill after strategy selection.

## Inputs
- State memory with derived flags from the state-memory skill.
- Candidate strategies containing `name`, `priority`, `predicate`, and `actions`.
- Optional previous strategy name.

## Outputs
- Selected strategy name, whether an interruption occurred, interruption reason, and explicit action queue.

## Workflow
1. Evaluate simple predicates against state flags.
2. Sort active strategies by descending priority and stable name.
3. Detect when the selected active strategy differs from the previous strategy.
4. Preserve contextual action wrappers as structured tokens, never ambiguous single keystrokes.
5. Return a safe wait action if no strategy is active.

## Validation
Run `python tests/test_controller.py` or validate the tree with `--run-tests`.

## Limitations
Predicates are deliberately simple for bounded recovery. The skill preserves AutoAscend's interruptible arbitration mechanism rather than full game play.

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.

Comments

Loading comments…

Newsletter

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