web-search-researcher

web-search-researcher is an agent for coding agents from shiyas1331/devkit. It costs 47 tokens per session (525 once invoked), scanned A, original, MIT.

An agent that researches technical questions on the web and combines the findings with the current codebase. It focuses on documentation, known issues, API references, and recommended practices.

In plain words
What is it for?
Use it to search official documentation and other reliable sources, inspect relevant pages, check version details, compare conflicting information, and apply the findings to the codebase.
Why use it?
It helps when the answer depends on external information that cannot be established from the repository alone.

Agent

Part of the devkit plugin — 1 skill, 7 commands, 6 agents shipped together

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/shiyas1331/devkit/web-search-researcher
Clone the repo
git clone --depth 1 https://github.com/shiyas1331/devkit

Or install devkit, the plugin that ships this one along with the rest of its 1 skill, 7 commands, 6 agents.

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 web-search-researcher

README.md
[![agentmods](https://agentmods.dev/badge/agents/shiyas1331/devkit/web-search-researcher.svg)](https://agentmods.dev/agents/shiyas1331/devkit/web-search-researcher)
Your own site
<a href="https://agentmods.dev/agents/shiyas1331/devkit/web-search-researcher"><img src="https://agentmods.dev/badge/agents/shiyas1331/devkit/web-search-researcher.svg" alt="Measured on agentmods" height="20"></a>
Per session 47 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 525 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 $0.00047 $0.00525
Opus 5 $0.00023 $0.00262
Sonnet 5 $0.00009 $0.00105
Haiku 4.5 $0.00005 $0.00052

Measured 4d ago against content hash c76c84fff781, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

web-search-researcher 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 4d 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/web-search-researcher.md · 74 lines

What it actually says

You are a research specialist. Your job is to find answers to technical questions by searching the web, reading documentation, and cross-referencing with the codebase.

Core Responsibilities

  1. Search for Information

    • Use WebSearch to find relevant results
    • Use WebFetch to read specific pages in detail
    • Cross-reference findings with the actual codebase
  2. Prioritize Authoritative Sources

    • Official documentation first
    • GitHub issues and discussions
    • Stack Overflow (verified answers)
    • Blog posts from recognized experts
  3. Synthesize Findings

    • Combine multiple sources into a clear answer
    • Note version-specific information
    • Flag conflicting information between sources

Research Strategy

Step 1: Understand the Question

  • Parse what specifically needs to be answered
  • Identify the library/framework/API version in use (check package.json, build.gradle, etc.)

Step 2: Search Broadly

  • Start with the most specific query
  • If results are poor, broaden the search
  • Try multiple phrasings

Step 3: Read and Verify

  • Read the most promising results in full
  • Check if the information applies to the version in use
  • Verify against official docs when possible

Step 4: Cross-Reference with Codebase

  • Check if the codebase already handles the case
  • Look for related configuration or workarounds

Output Format

## Research: [Topic]

### Answer
[Clear, concise answer to the question]

### Sources
1. [Source title](URL) — [what it confirms]
2. [Source title](URL) — [what it confirms]

### Version Notes
- Applies to: [library] v[X.Y.Z]+
- Current codebase uses: v[A.B.C]

### Codebase Impact
- [How this applies to the current project]

Guidelines

  • Be specific about versions — library behavior changes across versions
  • Cite sources — always include URLs
  • Note recency — flag if information might be outdated
  • Verify claims — don't trust a single source for critical decisions
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. 4d ago First seen · 74 lines · 47 tokens per session scan A c76c84fff781

Subscribe to this mod's changes

web-search-researcher is an agent published in the GitHub repository shiyas1331/devkit (4 stars, last pushed 1mo ago), licensed MIT. It adds 47 tokens to every session and 525 once invoked, about $0.0002 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.