Borrowing it
Nothing to install: this file belongs to maxwellsdm1867/wheeler. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/maxwellsdm1867/wheeler/main/.claude/skills/wheeler-service-creator/SKILL.mdgit clone --depth 1 https://github.com/maxwellsdm1867/wheelerWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/maxwellsdm1867/wheeler/wheeler-service-creator)<a href="https://agentmods.dev/skills/maxwellsdm1867/wheeler/wheeler-service-creator"><img src="https://agentmods.dev/badge/skills/maxwellsdm1867/wheeler/wheeler-service-creator/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/maxwellsdm1867/wheeler/wheeler-service-creator"><img src="https://agentmods.dev/badge/skills/maxwellsdm1867/wheeler/wheeler-service-creator.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 5.1 | $0.00346 | $0.09154 |
| Opus 5 | $0.00173 | $0.04577 |
| Sonnet 5 | $0.00069 | $0.01831 |
| Haiku 4.5 | $0.00035 | $0.00915 |
Grade A, and why
wheeler-service-creator 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 11d 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.
How it starts
The opening of the file, as written. The whole thing — 607 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Wheeler service creator
You scaffold a new Wheeler service adapter from a tool. A tool becomes a
declarative contract (one manifest entry the registry reads); a command opts
in via a flag; a run is one Execution whose wiring has THREE parts: structural
inputs (USED), structural outputs (WAS_GENERATED_BY), and SEMANTIC wiring of
the new outputs to the EXISTING graph (the Wheeler relationships, a judgment call
that lives in the act, not the parser). The Asta adapters are instance #1 and
your concrete template. You produce the four pieces, then hand the scientist a
short to-do list: capture one real output, fill the parser against it, run the
adversarial review, land it.
You write SKELETONS, not finished adapters. The parser is tool-specific and can
only be written against a real captured output, which you do not have at scaffold
time. Your job is to lay down every piece of the structure correctly so the human
fills one function (parse_<tool>) and reviews.
Two load-bearing ideas you must get right
Everything below is in service of these two. Read them first.
The registry reads the contract; the adapter does not hardcode the provider
The contract you write is DATA, read by wheeler/integrations/registry.py.
There are THREE distinct sets, do not conflate them:
- CATALOG (
wheeler/integrations/services.default.yaml): everything AVAILABLE to enable. Ships with the package.catalog_services()returns it. A service that should ship with Wheeler belongs here. - ENABLED set: what is actually LOADED and visible to the router and plan
acts. The source of truth is a FOLDER,
<project_root>/.wheeler/services/, with one<id>.yamlfile per enabled contract (same schema as a catalog entry: a bare mapping, or wrapped underservices:).load_services()returns this. Precedence: if the folder exists it IS the enabled set (an empty folder means nothing is enabled); else the legacy single-file.wheeler/services.yaml(backward-compat); else the full catalog (every default stays on until the user starts curating). - AVAILABLE: the enabled contracts whose
availableshell probe passes.available_services()filters the enabled set by the probe. This is what the router actually offers.
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 11d ago First seen · 607 lines · 346 tokens per session scan A 72b47994ee1a
wheeler-service-creator is a skill published in the GitHub repository maxwellsdm1867/wheeler (11 stars, last pushed 9d ago), licensed MIT. It adds 346 tokens to every session and 9,154 once invoked, about $0.0017 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.
Other skills, from other repositories
ml-paper-writing
Write publication-ready ML/AI papers for NeurIPS, ICML, ICLR, ACL, AAAI, COLM. Use when drafting papers from research repos, structuring arguments, verifying citations, or preparing camera-ready submissions. Includes LaTeX templates, reviewer guidelines, and citation verification workflows.
deepdive
A structured research workflow for investigating a question or decision using web and academic sources. It saves each source with quotations and details, then builds a question-and-answer report from reusable findings.
constitutional-ai
Anthropic's method for training harmless AI through self-improvement. Two-phase approach - supervised learning with self-critique/revision, then RLAIF (RL from AI Feedback). Use for safety alignment, reducing harmful outputs without human labels. Powers Claude's safety system.
consilium-connect
Onboarding funnel for Consilium-Principis — a personal board of AI personas of real thinkers (Sun Tzu, Marcus Aurelius, Epictetus, Machiavelli, plus your own legal figures) with a fail-closed fidelity contour, grounded in public-domain texts. This skill does NOT contain the advisor engine — that lives in a local MCP…
asta-skill
Domain expertise for Ai2 Asta MCP tools (Semantic Scholar corpus). Intent-to-tool routing, safe defaults, workflow patterns, and pitfall warnings for academic paper search, citation traversal, and author discovery.
consilium-principis
A decision-support tool that creates AI representations of selected thinkers and has them discuss a user’s question as a personal board of advisors.