Deploys an AI-driven NPC agent to perform game-world tasks. Uses a Pipecat/Gemini pipeline for autonomous behavior.
Install
mkdir -p .claude/skills/npc && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/17323" && unzip -o skill.zip -d .claude/skills/npc && rm skill.zipInstalls to .claude/skills/npc
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
Runs an autonomous AI task agent as a game character. Resolves a character name to its UUID, then launches the `npc-run` script which connects to the game server and executes the given task using a Pipecat + Gemini LLM pipeline.Key capabilities
- →Resolve character names to UUIDs
- →Launch the `npc-run` script for autonomous AI task agents
- →Execute tasks for game characters using Pipecat + Gemini LLM pipeline
- →Redirect all script output to a log file
- →Report character, task, PID, and log file path to the user
- →Provide instructions for checking progress and stopping the agent
How it works
The skill resolves a character's UUID from its name using Supabase, then launches the `npc-run` script in the background to execute a given task for the character using a Pipecat + Gemini LLM pipeline, redirecting all output to a log file.
Inputs & outputs
When to use npc
- →Assign a task to an NPC
- →Run an AI agent in a game environment
- →Automate game character behavior
About this skill
NPC
Runs an autonomous AI task agent as a game character. Resolves a character name to its UUID, then launches the npc-run script which connects to the game server and executes the given task using a Pipecat + Gemini LLM pipeline.
Parameters
- character_name (required): The character's display name (e.g.
JOETRADER). Passed as the argument to/npc. - task (required): A natural language description of what the character should do (e.g. "Explore and find 5 new sectors").
If character_name is not provided as an argument, ask the user for it. Always ask the user for the task description.
Steps
1. Source environment variables
set -a && source .env.supabase && set +a
2. Resolve the character UUID
Look up the character by name using Supabase REST:
curl -s "${SUPABASE_URL}/rest/v1/characters?name=eq.<character_name>&select=character_id,name" \
-H "apikey: ${SUPABASE_SERVICE_ROLE_KEY}" \
-H "Authorization: Bearer ${SUPABASE_SERVICE_ROLE_KEY}"
This returns a JSON array. Extract the character_id from the first result. If no results are returned, report that the character was not found and stop.
3. Run the NPC task agent
Run the npc-run script with the resolved character ID and the task. Redirect all output to a log file (do NOT use tee):
uv run npc-run <character_id> "<task>" > logs/npc-<character_name>.log 2>&1 &
NPC_PID=$!
echo "NPC agent started with PID $NPC_PID, logging to logs/npc-<character_name>.log"
Run the process in the background so it doesn't block the conversation.
4. Report to the user
Show the user:
- Character name and resolved UUID
- The task being executed
- The PID and log file path
- How to check progress:
tail -f logs/npc-<character_name>.log - How to stop the agent:
kill <PID>
Important notes
- Environment variables
GOOGLE_API_KEYandSUPABASE_URLmust be set (they come from.env.supabase). - Edge functions must be running:
npx supabase functions serve --workdir deployment --no-verify-jwt --env-file .env.supabase - The NPC agent runs autonomously and calls
finished()when the task is complete. - All output MUST go to a log file. Do NOT use
tee. Usetail,head, orgrepto inspect the log. - The character must already exist in the game and have joined (has a ship, is in a sector).
- For corporation ship control, the user can provide a
--ship-idflag manually if needed.
When not to use it
- →When the task does not involve an autonomous AI task agent as a game character
- →When the game server or Supabase environment is not set up
- →When the user needs real-time, interactive control over the NPC agent
Limitations
- →Environment variables `GOOGLE_API_KEY` and `SUPABASE_URL` must be set
- →Edge functions must be running for Supabase
- →All output MUST go to a log file; `tee` is not allowed
How it compares
This skill automates the deployment and task execution of AI-controlled game characters, providing a hands-off approach to managing NPC behavior, unlike manual control or direct scripting.
Compared to similar skills
npc side by side with the closest alternatives in the catalog.
| Skill | Installs | Updated | Safety | Difficulty |
|---|---|---|---|---|
| npc (this skill) | 0 | 5mo | Caution | Intermediate |
| computer-use-agents | 10 | 6mo | Review | Advanced |
| crewai-developer | 2 | 7mo | Review | Advanced |
| extra-rlhf | 0 | 1mo | Review | Advanced |
Try saying
Example prompts that trigger this skill in your AI assistant.
More by pipecat-ai
View all by pipecat-ai →You might also like
computer-use-agents
davila7
Build AI agents that interact with computers like humans do - viewing screens, moving cursors, clicking buttons, and typing text. Covers Anthropic's Computer Use, OpenAI's Operator/CUA, and open-source alternatives. Critical focus on sandboxing, security, and handling the unique challenges of vision-based control. Use when: computer use, desktop automation agent, screen control AI, vision-based agent, GUI automation.
crewai-developer
smallnest
Comprehensive CrewAI framework guide for building collaborative AI agent teams and structured workflows. Use when developing multi-agent systems with CrewAI, creating autonomous AI crews, orchestrating flows, implementing agents with roles and tools, or building production-ready AI automation. Essential for developers building intelligent agent systems, task automation, and complex AI workflows.
extra-rlhf
veceno
Use for anything about the ExtraArena RLHF data-collection & training-orchestration environment (rlhf_env, port 8090, MCP stdio): running semi-synthetic battles, generating training traces, orchestrating the Extra-LR training pipeline, or playing battles as a sub-agent. Routes to three sub-skills —
llama-cpp
zechenzhangAGI
Runs LLM inference on CPU, Apple Silicon, and consumer GPUs without NVIDIA hardware. Use for edge deployment, M1/M2/M3 Macs, AMD/Intel GPUs, or when CUDA is unavailable. Supports GGUF quantization (1.5-8 bit) for reduced memory and 4-10× speedup vs PyTorch on CPU.
langchain
zechenzhangAGI
Framework for building LLM-powered applications with agents, chains, and RAG. Supports multiple providers (OpenAI, Anthropic, Google), 500+ integrations, ReAct agents, tool calling, memory management, and vector store retrieval. Use for building chatbots, question-answering systems, autonomous agents, or RAG applications. Best for rapid prototyping and production deployments.
llama-factory
zechenzhangAGI
Expert guidance for fine-tuning LLMs with LLaMA-Factory - WebUI no-code, 100+ models, 2/3/4/5/6/8-bit QLoRA, multimodal support