Agent Building skills

Browse reusable Agent Skills, each with a clear purpose and practical guidance.

sft-launch

Launch SFT via `python -m hpc.launch --job_type sft` on any cluster (JSC Jupiter GH200, CINECA Leonardo A100, TACC Vista GH200), with EITHER backend — LLaMA-Factory (default) or axolotl (`--sft_backend axolotl`) — including Delphi tool-calling models (delphi template, tokenizer prep, jinja-as-ground-truth masking). This skill is the cluster-AGNOSTIC core (backend choice, Delphi handling, config maps, node-scaling, dataset mixing, cleanup recognition, common traps). Per-cluster particulars (preamble, paths, QOS/wall, sbatch patches, no-internet handling, HF-upload mechanics) live in `.claude/ops/<cluster>/ops.md §SFT`. Use when asked to SFT / launch a finetune / train a model on Jupiter, Leonardo, or TACC. Reference: notes/ot-agent/sft_experiments.md, CLAUDE.md.

259 repo starsObserved in 1 repos
Agent Building
GitHub

show-routine

Show the proposed sales routine without installing it — read-only preview of which frameworks YALC would auto-configure, with schedules and rationale, based on current archetype + providers + context. Use when the user says 'show my routine', 'what would YALC propose', 'preview the routine', 'dry-run the routine generator', or 'just show me the proposal'. Read-only — never writes anything.

259 repo starsObserved in 1 repos
Agent Building
GitHub

supervisor-init

Bootstrap the supervisor role at the start of a session — the human-facing lab supervisor who manages a large multi-experiment ML-ops operation: the single interface between the user and subagents / crons / top-level tools, keeper of secrets, and tracker of everything in flight. Run this FIRST in a fresh session (or when the user says "set up", "init", "take over", "you're the supervisor", "get oriented"). It walks an init checklist (orient in .claude, load the local env, take custody of secrets, survey in-flight work + crons + subagents), states the operating discipline (verify subagent work, fix proactively without gating unless needed, secrets only via env vars), and concludes by (re)creating the 3-hour sweep loop (monitor-restore) and running an initial sweep (monitor-cron-sweep).

259 repo starsObserved in 1 repos
Agent Building
GitHub

zicXYv2

This skill describes the `zicXYv2` project for AI agents working on code implementation and UI/audio behavior. The goal is to provide enough context so generated code matches the target platform, hardware, interaction model, and performance constraints.

259 repo starsObserved in 1 repos
Agent Building
GitHub

claw-advisor

This skill should be used when the user asks about OpenClaw configuration, troubleshooting, setup, architecture, or any OpenClaw question. Triggers on "how do I configure OpenClaw", "set up telegram in OpenClaw", "gateway configuration", "OpenClaw troubleshooting", "claw advisor", "what's the best way to set up OpenClaw", "OpenClaw docs", "help me with OpenClaw", "openclaw channel setup", "debug OpenClaw", or needs guidance on OpenClaw features, channels, gateway, automation, models, or design decisions.

258 repo starsObserved in 2 repos
Agent Building
GitHub

create-claw-skill

This skill should be used when the user asks to "create an OpenClaw skill", "make a claw skill", "build a skill for OpenClaw", "write a SKILL.md for openclaw", "add a skill to openclaw", "generate openclaw skill frontmatter", "create a clawhub skill", "port a skill to OpenClaw", "convert a Claude Code skill to claw", "migrate my skill to openclaw", or wants to author a new skill or port an existing Claude Code skill for the pi-coding-agent / OpenClaw ecosystem.

258 repo starsObserved in 2 repos
Agent Building
GitHub

agentic-clarification

Use when a user's request is vague, ambiguous, or underspecified. Runs recon-first clarification — codebase exploration before questions, then one bundled question round with defensible defaults — and outputs a Goal Contract for the durable /goal runtime. Triggers on "I want to...", "I need...", "let's build...", "can you help me...", "we should...", or any request where the full scope isn't immediately clear.

258 repo starsObserved in 1 repos
Agent Building
GitHub

agentic-goal

Primary execution workflow for durable /goal runtime. Use when a Goal Contract is active or when the user asks to execute, continue, verify, or complete a goal.

258 repo starsObserved in 1 repos
Agent Building
GitHub

get-token-insights

Use this skill when the user wants to analyze Claude token usage, understand Claude API spending, check cache hit rates, review Claude Code workflow patterns (skills, agents, hooks), or get cost optimization recommendations.

258 repo starsObserved in 1 repos
Agent Building
GitHub

repair-agent

This skill should be used when the user asks to "repair an agent", "audit an agent", "fix my agent", "review agent quality", "check if my agent is well-written", "diagnose agent problems", "what's wrong with this agent", "improve this agent", or "what's wrong with this agent file". Not for skills — use repair-skill.

258 repo starsObserved in 1 repos
Agent Building
GitHub