skill-eval-improve
Improves Agent Skills via validate → rule-based eval cases → plugin-eval → prompt evals → bounded edits with held-out gates. Use when tuning skill quality,…
Find an Agent Skill for your task. Read the guide and copy the install command.
Updated 11-05 00:16
25,653 results
Improves Agent Skills via validate → rule-based eval cases → plugin-eval → prompt evals → bounded edits with held-out gates. Use when tuning skill quality,…
Builds apps on top of cognee v0.5.8, the knowledge-graph memory engine for AI agents. Use when ingesting text/files/URLs into persistent agent memory, building…
Generate the complete AscendC operator code implementation based on the design document and complete framework adaptation. TRIGGER when: the design document is complete and op_host/op_kernel code needs to be generated, registered to the PyTorch framework, and compiled and tested. Keywords: code generation, op_host, op_kernel, tiling, kernel, framework adaptation, operator registration.
Verl single-async DAPO training configuration generator. Trigger scenarios: (1) Start single-async DAPO training (2) Generate training script (3) Configure feature parameters (4) Pre-training check. **Feature strategy**: When the user does not specify, enable performance features by default (flash_attn/dynamic_batch/remove_padding/gradient_checkpoint...
One-click deployment of vLLM large model inference services on the Ascend NPU platform. Trigger: the user says 'deploy model name', 'NPU deploy model', 'vllm serve'. Process: SSH check → NPU check → configuration discovery (must be verified) → user confirmation → deployment → cron monitoring → verification. Constraints: (1) Configuration must be verified from official documentation; guessing is prohibited; (2)…
Universal MindSpeed-MM generative model training guide for Huawei Ascend NPU. Covers all backend patterns (Megatron, Megatron+FSDP2, FSDP2-native,…
Universal VLM (vision-language understanding model) training guide for Huawei Ascend NPU using MindSpeed-MM. Covers all three framework patterns (Megatron,…
Run a goal as a project: you coordinate. Use for "agents <task>", "take this to done", "work on this in parallel", "what are my threads doing", "pick up…
Create and validate a new native Muse plugin package in the current workspace. Use ONLY when the user explicitly asks to create a Muse plugin or invokes the…
Create and validate a new Muse skill — project-local in the current workspace by default, or a personal skill staged for `muse skills install` into the managed…
Read and create notes through Evernote's official MCP server.
Coding-agent sessions on every machine you connected, over Herdr where it runs, tmux elsewhere, and MSP for your MSP hosts: one digest of what is waiting on…
Search and work with the user's GitHub repositories through GitHub's official MCP server.
Bring what Claude Code or Codex already remembers or has configured for this user into Muse Code — their memory notes and their MCP servers — whenever the user…
Read and manage the user's QuickBooks business through Intuit's official MCP server, including reports, invoices, customers, products, payment links, sales…
Locate and read Muse Code's OWN session logs — the current session or a prior one. Use when the user asks to pull context from, continue, summarize, or inspect…
Continue work from a local Claude Code session in Muse Code. Use when the user asks to resume, continue, or recover unfinished Claude Code work, with or…
Continue work from a local Codex session in Muse Code. Use when the user asks to resume, continue, or recover unfinished Codex work, with or without a session…
Ground self-referential answers in the agent's actual filesystem. Use when the user asks who the agent is, what it can do, what it knows, what it remembers,…
Publish a reusable native Muse Idea with idea.share when the user explicitly asks to share or publish an Idea or asks for a muse.ai/ideas link. Never offer it…
Set up and operate a Shopify store with Shopify's official MCP server (the installed `shopify` command, NOT the npm Shopify CLI). Use when someone wants to…
Keep a Markdown table readable in a narrow terminal of about 100 display columns - a wide table or one carrying prose in its cells wraps into unreadable ragged…
Read and manage Todoist tasks, projects, comments, labels, filters, and reminders through Todoist's official MCP server.
Provides the static system voice catalog used by Jarvis. It does not define a user-facing workflow.