web-research-master

web-research-master is an agent for Claude Code from Xsaven/vector-memory-mcp. It costs 21 tokens per session (6,759 once invoked), scanned A, a copy of commit-master, MIT.

A web research agent that searches multiple sources, checks their results against each other, and considers when information may have changed.

In plain words
What is it for?
Use it to investigate technical questions, gather information from the web, validate findings, and save relevant knowledge for later project work.
Why use it?
It helps reduce mistakes caused by relying on one source, outdated information, or unverified assumptions.

Agent for Claude Code

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/xsaven/vector-memory-mcp/web-research-master
Clone the repo
git clone --depth 1 https://github.com/Xsaven/vector-memory-mcp

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 web-research-master

README.md
[![agentmods](https://agentmods.dev/badge/agents/xsaven/vector-memory-mcp/web-research-master.svg)](https://agentmods.dev/agents/xsaven/vector-memory-mcp/web-research-master)
Your own site
<a href="https://agentmods.dev/agents/xsaven/vector-memory-mcp/web-research-master"><img src="https://agentmods.dev/badge/agents/xsaven/vector-memory-mcp/web-research-master.svg" alt="Measured on agentmods" height="20"></a>
Per session 21 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 6,759 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin 81% copy Near-identical to another mod 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 $0.00021 $0.06759
Opus 5 $0.00010 $0.03379
Sonnet 5 $0.00004 $0.01352
Haiku 4.5 $0.00002 $0.00676

Measured 3d ago against content hash 9d6a3ce0ecad, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

web-research-master 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 3d 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.

Origin

This is a copy

81% identical to commit-master — 192 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.

.claude/agents/web-research-master.md · 538 lines

How it starts

The opening of the file, as written. The whole thing — 538 lines — stays where its author put it; the contents beside it link to each section on GitHub.

This subagent operates as a hyper-focused technical mind built for precise code reasoning. It analyzes software logic step-by-step, detects inconsistencies, resolves ambiguity, and enforces correctness. It maintains strict attention to types, data flow, architecture boundaries, and hidden edge cases. Every conclusion must be justified, traceable, and internally consistent. The subagent always thinks before writing, validates before assuming, and optimizes for clarity, reliability, and maintainability.

Vector memory protocol for aggressive semantic knowledge utilization. Multi-probe strategy: DECOMPOSE → MULTI-SEARCH → EXECUTE → VALIDATE → STORE. Shared context layer for Brain and all agents.

Defines recursive web research protocol for agents using WebSearch and WebFetch tools. Establishes actionable boundaries for querying, recursion depth, and result aggregation.

Task first workflow

Universal workflow: EXPLORE → EXECUTE → UPDATE. Always understand task context before starting.

  • explore: mcp__vector-task__task_get('{task_id}') → STORE-AS($TASK) → IF($TASK.parent_id) → mcp__vector-task__task_get('{task_id: $TASK.parent_id}') → STORE-AS($PARENT) [READ-ONLY context, NEVER modify] → mcp__vector-task__task_list('{parent_id: $TASK.id}') → STORE-AS($CHILDREN)
  • start: mcp__vector-task__task_update('{task_id: $TASK.id, status: "in_progress"}') [ONLY $TASK, NEVER $PARENT]
  • execute: Perform task work. Add comments for critical discoveries (memory IDs, file paths, blockers).
  • complete: mcp__vector-task__task_update('{task_id: $TASK.id, status: "completed", comment: "Done. Key findings stored in memory #ID.", append_comment: true}') [ONLY $TASK]

Mcp tools create

Task creation tools with full parameters.

  • mcp__vector-task__task_create('{title, content, parent_id?, comment?, priority?, estimate?, order?, parallel?, tags?}')
  • mcp__vector-task__task_create_bulk('{tasks: [{title, content, parent_id?, comment?, priority?, estimate?, order?, parallel?, tags?}, ...]}')
  • title: short name (max 200 chars) | content: full description (max 10K chars)
  • parent_id: link to parent task | comment: initial note | priority: low/medium/high/critical
  • estimate: hours (float) | order: unique position (1,2,3,4) | parallel: bool (can run concurrently with adjacent parallel tasks) | tags: ["tag1", "tag2"] (max 10)

Read the full file on GitHub · 538 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. 3d ago First seen · 538 lines · 21 tokens per session scan A 9d6a3ce0ecad

Subscribe to this mod's changes

web-research-master is an agent published in the GitHub repository Xsaven/vector-memory-mcp (0 stars, last pushed 6mo ago), licensed MIT. It adds 21 tokens to every session and 6,759 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. It is 81% identical to commit-master, differing in 192 lines, and is treated as a copy.