Search

A first-person cave-horror game for the terminal, with scenes raymarched at runtime and lit by your own lamp.

Administer, score and record validated personality and affect questionnaires.

A wrapper around bullmq

Claude Code skill: Fable 5.1 decides, Opus 5 / Sonnet 5 / Haiku 4.5 execute. Five wiring patterns (advisor, architect+delegate, agent teams, workflows, /goal), effort routing, fallback traps, 16-block prompt kit, pinned agent lanes.

Check whether a website is reachable and readable by AI crawlers, and score it. Two scores, never averaged. Blocking model training costs no points.

A bounded Riftbound rules engine and AI-agent toolkit: Deck Coach, Rule Consult, and Player2 infrastructure with versioned grammar and re-runnable evidence for covered mechanics. Not a game simulator or official rules authority. Licensed card programs live in a private overlay. Unofficial fan project.

Research Software Engineering skills for AI coding agents

A reusable research-onboarding skill for Claude, ChatGPT, Codex, and other instruction-following agents.

Agent skill for Claude Code, Codex and Cursor: generate or refresh project rules (.cursor/rules/*.mdc, CLAUDE.md and .claude/rules, a .codex hook bridge that attaches Cursor rules by glob) with layered-cake architecture, BDD delivery and a mandatory Rules Sync step across every agent tree

Runaway tool-call argument guard for dsh: wrap the llm/stream waterfall and cut a tool call that streams more than a configurable number of arguments bytes, terminating with a routed TOOL_CALL_ARGUMENTS_TOO_LARGE error finish so the oversized call is neve

SUPlime speaker diarization: WavLM / Conformer segmentation + SimAM-ResNet34 embeddings, loadable through pyannote.audio 4.x

Boruna: a deterministic, capability-safe programming language and framework for LLM-native applications