add-llama-cpp

A setup guide for running llama.cpp, a local server for generating text with AI models, and optionally connecting it to Deus.

In plain words
What is it for?
Installing and checking llama.cpp on macOS, starting its local server, and connecting it to Deus when the optional llama_cpp provider is already supported.
Why use it?
It provides a possible faster or lower-cost local text-generation route for experiments. It does not replace Ollama for embeddings or the default judging tasks.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/sliamh11/deus/add-llama-cpp
Any agent
npx skills add sliamh11/Deus --skill add-llama-cpp
Clone the repo
git clone --depth 1 https://github.com/sliamh11/Deus

Made for: Claude Code, Codex.

Per session 35 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,585 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00035 $0.02585
Opus 5 $0.00017 $0.01293
Sonnet 5 $0.00007 $0.00517
Haiku 4.5 $0.00003 $0.00259

Measured 2d ago against content hash 8eec98df2db6, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

add-llama-cpp scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

curl -fsS http://127.0.0.1:8080/health 2>/dev/null || echo "llama-server not responding on 127.0.0.1:8080"
.claude/skills/add-llama-cpp/SKILL.md · 282 lines

How it starts

The opening of the file, as written. The whole thing — 282 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Add llama.cpp

This skill installs llama.cpp, runs llama-server as a local host service, and wires the local endpoint into Deus only when the current checkout already supports the optional llama_cpp provider.

Use this when the user wants a faster or cheaper local text-generation path for experiments, benchmarks, or future backend work.

Slash command: /add-llama-cpp.

Important boundaries:

  • This does not replace Ollama for memory embeddings or the default judge. Ollama remains required unless the repo deliberately changes those surfaces.
  • This skill is macOS-first for installation and service management. On Linux or Windows, continue only if llama-server is already installed or the user explicitly wants a manual install path.
  • If the current checkout does not yet contain the optional Deus-side llama_cpp integration, complete the host install anyway and tell the user the runtime wiring is a separate source task.

Phase 1: Pre-flight

Check current state

command -v llama-server >/dev/null 2>&1 && llama-server --version || echo "llama.cpp not installed"
curl -fsS http://127.0.0.1:8080/health 2>/dev/null || echo "llama-server not responding on 127.0.0.1:8080"
test -f evolution/generative/providers/llama_cpp.py && echo "DEUS_LLAMA_CPP_PROVIDER=true" || echo "DEUS_LLAMA_CPP_PROVIDER=false"
test -f setup/llama-cpp.ts && echo "DEUS_LLAMA_CPP_SETUP=true" || echo "DEUS_LLAMA_CPP_SETUP=false"

Ask scope

AskUserQuestion: Do you want host install only, or host install plus optional Deus wiring if this checkout supports it?

  1. Host install only - install llama.cpp, run llama-server, and verify the local endpoint
  2. Host install + Deus wiring - also configure repo env vars and run Deus-side verification when the checkout supports it

If the user chooses option 2 but either provider/setup file is missing, say clearly that the host install can proceed now and the checkout wiring remains a separate code task.

Phase 2: Install llama.cpp

Read the full file on GitHub · 282 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 282 lines · 35 tokens per session scan A 8eec98df2db6

Subscribe to this mod's changes

add-llama-cpp is a skill published in the GitHub repository sliamh11/Deus (51 stars, last pushed 8d ago), licensed MIT. It adds 35 tokens to every session and 2,585 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories