Oracle is a command-line tool and MCP server that packages a prompt with selected project files, sends them to an AI model, and saves the response as a session. Developers and coding agents use it to get a second-model review based on the actual code and context. The catalogue add-ons support Oracle workflows.
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.
npx skills add steipete/oracle --skill oracle-advisorgit clone --depth 1 https://github.com/steipete/oracleWrote 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/steipete/oracle/oracle-advisor)<a href="https://agentmods.dev/skills/steipete/oracle/oracle-advisor"><img src="https://agentmods.dev/badge/skills/steipete/oracle/oracle-advisor/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/steipete/oracle/oracle-advisor"><img src="https://agentmods.dev/badge/skills/steipete/oracle/oracle-advisor.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00049 | $0.00830 |
| Opus 5.5 | $0.00020 | $0.00332 |
| Sonnet 5.5 | $0.00010 | $0.00166 |
| Haiku 4.5 | $0.00005 | $0.00083 |
Grade A, and why
oracle-advisor 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 1mo 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 — 78 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Oracle advisor
Use Oracle to ask an advisory model about selected context. The advisor returns advice; the calling agent remains responsible for checking it and making edits. This skill reuses the existing Oracle CLI and MCP tools. It does not add a new command, MCP method, or native desktop adapter.
Select the allowed transport
Honor the caller's transport, model, effort, files, output requirements, and choice of a fresh consultation or an explicit follow-up. Existing authorization persists; do not ask again when a permitted route is already clear.
- API: use an explicitly configured and authorized provider. Pass
--engine apiand the requested model; preserve required reasoning settings. - Browser: use the authorized signed-in browser route. Pass
--engine browser; confirm the requested model and effort through Oracle's selection evidence. Browser availability alone does not authorize a fallback. - Render: use
--renderto prepare a bundle for manual handoff. This creates no model answer or completed consultation; report the bundle as ready to send. - Native desktop: unavailable in this skill. Chat, Work, and Codex sharing a UI is not evidence of an agent-callable delegation interface. Do not improvise private IPC, cookie extraction, or UI automation as a native adapter.
If the requested route is unavailable, use an alternative only when the caller has already allowed it. Otherwise explain the missing capability and ask for the transport choice. Never silently reduce the required model or effort.
Run the consultation
Inspect oracle --help --verbose for the installed version. For an execution
route, preview the exact bundle with the chosen engine and --dry-run full;
for render, inspect the rendered bundle directly. Use --files-report for token
estimates. A preview's predicted route is not execution evidence. Include the problem,
relevant constraints, and the requested answer format in the prompt. Send only
the selected task context, with credentials excluded.
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.
- 1mo ago First seen · 78 lines · 49 tokens per session scan A b2bad1497e2e
oracle-advisor is a skill published in the GitHub repository steipete/oracle (4,059 stars, last pushed yesterday), licensed MIT. It adds 49 tokens to every session and 830 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-09-08.
Other skills, from other repositories
generative-ui
Use when presenting information already in hand as a small, static structured panel in the conversation — status, results, comparisons, or a compact summary — rather than prose, a markdown table, a widget, or an app.
st-widget-builder
Use when the user asks to render, build, visualise, mock up, or show a widget (chart, dashboard, card, table, graph) in the Mewbo Console. Teaches how to delegate the work to the st-widget-builder sub-agent instead of writing HTML or inline code yourself.
device-control
Use when driving the user's Android device through devicecontrolstart, deviceui and deviceaction — opening apps, navigating screens, filling fields, or any multi-step flow through another app's UI.
mewbo-harness
Use when you need to know how this harness itself behaves rather than what a tool's arguments are — why a tool result was truncated and how to page past it, what a tool timeout does and does not mean, which directory a relative path resolves against, how to reach a tool whose schema was not bound, how spawned…
brainstorming
Explore intent, requirements and design before implementation. Use before any creative work such as creating features, building components, adding functionality, or modifying behavior.
claude-md-improver
Audit and improve CLAUDE.md files in repositories. Use when asked to check, audit, update, improve, or fix CLAUDE.md files, or when someone mentions project memory optimization.