A setup guide for xllm, a tool that lets different AI model providers act as coding advisors for a project.

Use it to inspect the machine, configure project settings, test live provider access, and save the chosen advisor setup.

It checks which advisor programs and local models are available, then helps choose a provider, model, and effort level for each role.

setup — Skill by kimmingul

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/kimmingul/xllm/setup

Any agent

npx skills add kimmingul/xllm --skill setup

Clone the repo

git clone --depth 1 https://github.com/kimmingul/xllm

Made for: Claude Code, Codex.

Per session 87 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 1,003 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. 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.

Model	Per session	Once invoked
Fable 5	$0.00087	$0.01003
Opus 5	$0.00044	$0.00502
Sonnet 5	$0.00017	$0.00201
Haiku 4.5	$0.00009	$0.00100

Measured 2d ago against content hash 20bf536d77d0, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

setup scanned grade A with 0 findings 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.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/setup/SKILL.md · 98 lines

How it starts

The opening of the file, as written. The whole thing — 98 lines — stays where its author put it; the contents beside it link to each section on GitHub.

setup — Machine inventory + per-project advisor wizard (xllm)

Resolve the advisor script

Claude Code: "${CLAUDE_PLUGIN_ROOT}/scripts/xllm-advisor.js". Codex / other hosts: the plugin root is two directories above this SKILL.md — use <plugin-root>/scripts/xllm-advisor.js.

Step 1 — Machine inventory (what CAN run here)

node <advisor.js> --inventory            # cached (24h TTL)
node <advisor.js> --inventory --refresh  # force re-probe

Reports per provider: installed, healthy, tier (strong/balanced/local), relative cost, and for ollama the actually pulled models. Cloud model catalogs are not enumerated — for cloud CLIs, installed means the binary responds; auth is only proven by node <plugin-root>/scripts/smoke.mjs --live.

Step 2 — Project marker + artifact dirs

node <advisor.js> --remember

Writes .xllm/xllm-advisor-path (legacy .grok/ honored) and creates secret-redacting artifact dirs with a self-ignoring .gitignore.

Step 3 — Per-project advisor wizard (posture packs)

Resolve pins deterministically; the skill only renders and confirms.

Preview the recommended pack (default balanced):
```
node <advisor.js> --setup balanced --json
```
The resolver returns { roles, warnings, evidence, recommended_packs }. balanced leaves analysis/design OPEN (measured routing) and pins at most a free local critic — a pin FREEZES measured routing, so packs pin only genuine constraints. quality = max-spend lock, frugal = cost lock, local = offline lock, skip = clear pins.
Ask ONE question using the host's UI, offering the first four of recommended_packs (always include skip). On Claude Code use AskUserQuestion with the resolver's top pack labeled "(Recommended)"; show the effort legend (Quick=low / Standard=medium / Deep=high) and one-line role glosses. Never invent cloud model names — cloud pins omit the model.
Show the resolved preview (roles + warnings + which stay OPEN and why), then on the user's accept:

Read the full file on GitHub · 98 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.

2d ago First seen · 98 lines · 87 tokens per session scan A 20bf536d77d0

Subscribe to this mod's changes

setup is a skill published in the GitHub repository kimmingul/xllm (2 stars, last pushed 9d ago), licensed MIT. It adds 87 tokens to every session and 1,003 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Data and AI Debugging Testing claude-code-plugin codex-plugin grok-build-plugin

watch

File sentinel that monitors the working directory for changes and marker comments, then auto-triggers appropriate skills. Poll-based via git diff against the last scan commit. Writes intake items for batch processing and routes marker actions through /do. Use for automatic reactions to file changes; do NOT use for…

SethGammon/Citadel · 70 tokens

review

5-pass structured code review — correctness, security, performance, readability, consistency.

SethGammon/Citadel · 17 tokens

live-preview

Mid-build visual verification loop. Takes screenshots of components during construction, not just after. Catches visual regressions and invisible features before they compound. Requires Playwright or similar screenshot tool.

SethGammon/Citadel · 40 tokens

marshal

Meta-orchestrator that takes any direction — broad, specific, or vague — and autonomously chains skills and context into actionable work. Gathers context from codebase, docs, and memory. Only asks the user when it genuinely cannot proceed. Single-session orchestrator.

SethGammon/Citadel · 56 tokens

wiki

Markdown-first knowledge base where the LLM acts as librarian. Ingests raw sources, compiles and interlinks topic files, self-maintains an index. No vector DB or embeddings required -- uses LLM-native navigation over structured markdown up to 400K words.