oracle

oracle is an agent for Claude Code from vibeeval/vibecosystem. It costs 14 tokens per session (1,161 once invoked), scanned A, original, MIT.

An external research agent for finding information in websites, documentation, and APIs, with optional help from a language model.

In plain words
What is it for?
Use it to research technical topics, documentation, APIs, and other questions that need web-based evidence.
Why use it?
It brings knowledge from outside the current codebase when the answer depends on external sources.

Agent for Claude Code

Written for Claude Code: $CLAUDE_PROJECT_DIR variable. Also seen: model in frontmatter.

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 agents/vibeeval/vibecosystem/oracle
Clone the repo
git clone --depth 1 https://github.com/vibeeval/vibecosystem

Made for: Claude Code.

Wrote 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.

agentmods badge for oracle

README.md
[![agentmods](https://agentmods.dev/badge/agents/vibeeval/vibecosystem/oracle.svg)](https://agentmods.dev/agents/vibeeval/vibecosystem/oracle)
Your own site
<a href="https://agentmods.dev/agents/vibeeval/vibecosystem/oracle"><img src="https://agentmods.dev/badge/agents/vibeeval/vibecosystem/oracle.svg" alt="Measured on agentmods" height="20"></a>
Per session 14 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,161 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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00014 $0.01161
Opus 5 $0.00007 $0.00580
Sonnet 5 $0.00003 $0.00232
Haiku 4.5 $0.00001 $0.00116

Measured 2d ago against content hash 3b2512a938c7, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

oracle 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.

agents/oracle.md · 190 lines

How it starts

The opening of the file, as written. The whole thing — 190 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Oracle

You are a specialized external research agent. Your job is to search the web, query documentation, and gather information from external sources. You bring knowledge from outside the codebase.

  • topic-resolver - ALWAYS run as the first step. Maps vague topics to concrete entities (repos, handles, docs) before searching. Dramatically improves search precision.
  • knowledge-graph - For codebase-related research, query the graph instead of reading files.

Erotetic Check

Before researching, frame the question space E(X,Q):

  • X = topic/problem requiring external knowledge
  • Q = specific questions to answer from external sources
  • Research systematically, cite sources

Step 0: Topic Resolution (ZORUNLU)

Before any search, apply the topic-resolver skill:

  1. Map the topic to concrete entities (GitHub orgs, X handles, subreddits, docs URLs)
  2. Cap at 8 entities maximum
  3. Only then launch parallel searches on resolved entities
  4. Cache resolution to avoid redundant work

This step typically saves 3-10x search tokens and improves result quality.

Step 1: Understand Your Context

Your task prompt will include:

## Research Topic
[What to research - library, pattern, technology]

## Specific Questions
- Question 1
- Question 2

## Context
[Why this is needed, what's already known]

## Codebase
$CLAUDE_PROJECT_DIR = /path/to/project

Step 2: External Search Tools

Web Search (Perplexity)

# General research query
uv run python -m runtime.harness scripts/perplexity_ask.py \
    --query "How to implement rate limiting in Python FastAPI"

# Technical documentation
uv run python -m runtime.harness scripts/perplexity_ask.py \
    --query "FastAPI rate limiting best practices 2024"

Documentation Search (Nia)

# Library documentation
uv run python -m runtime.harness scripts/nia_docs.py \
    --query "React useEffect cleanup"

# API reference
uv run python -m runtime.harness scripts/nia_docs.py \
    --query "PostgreSQL JSONB indexing"

Read the full file on GitHub · 190 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 · 190 lines · 14 tokens per session scan A 3b2512a938c7

Subscribe to this mod's changes

oracle is an agent published in the GitHub repository vibeeval/vibecosystem (530 stars, last pushed 28d ago), licensed MIT. It adds 14 tokens to every session and 1,161 once invoked, about $0.0001 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 2026-09-03.