krystophny
30 verified skills30 total stars
gpt
Delegate a task to GPT-5.4 high via one-shot codex exec. Use when the current local model needs a more capable model for a hard subtask.
development1
autonomous
Execute tasks using Claude Code CLI in autonomous mode. Handles single prompts and multi-step TODO lists with iterative review, commit, and push. Use when running autonomous Claude execution from Codex CLI.
tools1
beat-orchestrator
Use on every GM response during a Daggerheart session. Pipeline sequencing — mode detection, role dispatch, checkpoint enforcement, scene discipline, roll gate. Works alongside the causal-auditor skill.
testing1
check
Review changes before commit. Supports multiple PRs in parallel via worktrees. No args = current changes or PR.
testing1
commit
Create a git commit following CLAUDE.md rules. Verifies tests pass and stages files explicitly.
testing1
daggerheart-gm
RETIRED — Replaced by beat-orchestrator + causal-auditor skills. Do not invoke. Kept for reference only.
testing1
debug
Debug failing tests. Supports multiple tests in parallel. No args = all failing tests.
development1
do
Enforce CLAUDE.md framework compliance. Re-reads rules, displays applicable rules, processes request with full compliance.
development1
fix
Fix failing PR CI checks. Supports multiple PRs in parallel via worktrees. No args = current PR or create new.
testing1
gm-spatial-temporal-stress-test
Automated spatial-temporal consistency testing aligned with the 4-role, 3-checkpoint engine. Dispatches chaos agents to derail scenes and GM subagents to narrate with explicit State Gate and post-prose validation, then audits trackers for teleportation, drift, checkpoint misses, scale fidelity, and compass anchor compliance across 5-beat mini-sessions. Never invoke during play sessions.
testing1
gm-stress-test
Automated GM consistency, calibration, and line enforcement testing. Dispatches chaos player agents and GM subagents against a scenario bank, then audits rulings. Never invoke during play sessions.
testing1
helpy
Drive the helpy CLI for live office/web automation — SearXNG web search/fetch, TU Graz TUGonline (lectures, rooms, exams, candidates, Lehrerhebung), ICS feed subscription management, xlsx via excelize (read/write/formulas/recalc/charts/pivots/validation), Nextcloud shares and version history, SAP budget/OData reports, and read-only ops scans. Use whenever the task needs the live web, TUGonline, spreadsheets, ICS feeds, Nextcloud, or SAP. Do not use for generic mail/calendar; use sloptools for those.
tools1
multisprint
Continuous sprint cycles. Infinite WORK, PLAY, PLAN loops until user interruption. Use for extended autonomous development sessions.
development1
office-snapshot
One-shot local office snapshot across work/private mail, calendar, and web search. Use when the user asks Pi to show recent emails, recent calendar entries, search the web, make a daily brief, benchmark tool use, or combine sloppy/sloptools and helpy in one compact response.
tools1
opencode
Delegate a bulk-work subtask to the local Qwen via one-shot opencode run. Use when the subtask is high-volume but low-complexity (file scans, log parsing, large-text summaries, repetitive transforms) so it should not burn parent-model tokens.
development1
pi
Delegate a bulk-work subtask to the local Qwen via one-shot pi run. Use when the subtask is high-volume but low-complexity (file scans, log parsing, large-text summaries, repetitive transforms) so it should not burn parent-model tokens.
data-ai1
pr
Start new bug fix. Supports multiple in parallel via worktrees. No args = create from current changes.
tools1
slopask
Manage anonymous Q&A rooms for university lectures via the slopask CLI (deployed at ask.sloppy.at, SQLite + go-chi, no tracking). Use to create/list/close lecture Q&A rooms, mint admin and student URLs, or operate the slopask service.
tools1
slopcast
Render lecture videos from Beamer PDFs and pdfpc speaker notes via the slopcast pipeline (slide.pdf → segments.json → audio → clips → final/lecture.mp4) with Qwen3-TTS voice cloning and optional split-screen livecode. Use to generate, re-render, or splice slopcast lecture videos, or to author Beamer notes for slopcast.
tools1
sloptools
The user's email, contacts, personal tasks/todos, and full-CRUD Google + EWS calendars. Drives the sloptools CLI (same surface as the sloppy MCP on 127.0.0.1:9420). Use for mail (Gmail / Exchange-EWS / IMAP — list, read, send, reply, forward, flag, categorize, server-side filters, delegated mailboxes, out-of-office), calendar events (create / update / delete / RSVP / freebusy / ICS export across work + private accounts), contacts and contact groups, tasks (Google Tasks, Todoist), slopshell canvas, agent handoffs, and workspace items/artifacts/actors/triage.
tools1
sprint
Complete sprint cycle executing WORK, PLAY, PLAN phases. Autonomous batch execution with CI validation. Use to run a full sprint workflow.
testing1
sprint-plan
Strategic sprint planning via chris-architect. Issue consolidation, GitHub meta-issue updates, and sprint goal establishment. Use after PLAY phase or when planning next sprint.
development1
storyteller-disposition
RETIRED — Content moved to vault GM/Storyteller-Disposition.md. Use RAG query: query('storyteller disposition sensory mode editorial tendency') to find it.
data-ai1
work
Sprint implementation phase. Complete SPRINT BACKLOG items via max, implementer, reviewer, max agent chain with CI validation. Use for implementation work in sprint cycles.
testing1
comments
Address PR comments. Supports multiple PRs in parallel via worktrees. No args = current PR.
documentation1
codex
Execute tasks by invoking the codex agent. Handles single prompts and multi-step task documents. Use for stepwise Codex orchestration with autonomous execution.
development1
play
Sprint review and defect discovery. Sequential Task tool delegation through max, patrick, vicky, chris agents. Files GitHub issues for all discovered defects. NO file modifications.
tools1
issue
Create a properly formatted GitHub issue with MRE and version info.
development1
slopvault
ETL pipeline that imports manually-downloaded Discord, LinkedIn, and WhatsApp archive ZIPs into the user's brain vaults as plain files (no APIs, no tokens, no daemons). Use when the task involves processing or querying a Discord/LinkedIn/WhatsApp data export.
development1
causal-auditor
Use on every GM response during a Daggerheart session. Always-on validation — State Gate, checkpoint gates, entity census, drift check, tracker discipline. Works alongside the beat-orchestrator skill.
testing1