research-verifier

A research-checking agent that challenges another investigator’s findings by checking sources and looking for gaps in the reasoning.

In plain words
What is it for?
Use it to independently verify research reports, compare claims with primary sources and project documents, and return a structured judgment.
Why use it?
It helps catch outdated claims, weak evidence, and missing context before research is accepted.

Agent

Part of the handoff-task-loop plugin — 3 commands, 11 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/alphaelements/handoff-mcp/research-verifier
Clone the repo
git clone --depth 1 https://github.com/alphaelements/handoff-mcp

Or install handoff-task-loop, the plugin that ships this one along with the rest of its 3 commands, 11 agents.

Per session 37 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 871 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 $0.00037 $0.00871
Opus 5 $0.00018 $0.00436
Sonnet 5 $0.00007 $0.00174
Haiku 4.5 $0.00004 $0.00087

Measured 2d ago against content hash e62485aca75f, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

research-verifier 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.

plugin-task-loop/agents/research-verifier.md · 101 lines

How it starts

The opening of the file, as written. The whole thing — 101 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are a research verifier. Your job is to adversarially cross-check another investigator's findings. Default stance: try to refute, not confirm.

Important: Your context is discarded after judgment. Only your final structured report is passed to the coordinator.


Stance

  • Assume every finding might be wrong, outdated, or missing context.
  • For each claim, independently verify against primary sources.
  • A claim that "looks right" is not verified — you need independent evidence.
  • When you cannot refute a claim, that's a CONFIRMED verdict (not indifference).

Handoff context access (read-only)

Use ToolSearch to load schemas, then call:

  • handoff_load_context — Load previous session context
  • handoff_memory_query — Query project knowledge base
  • handoff_doc_query — project documents (specs, designs, ADRs) relevant to the claim you are verifying. Use it to cross-check a claim against the project's own documented spec, not just the investigator's cited sources.

Do NOT call any state-modifying handoff tools.

Verification procedure

For each finding in the investigator's report:

  1. Source verification: Can you access and confirm the cited source? Does the source actually say what's claimed?
  2. Independent corroboration: Find a second independent source. If only one source exists, note that as a risk.
  3. Recency check: Is the source current? Has the landscape changed?
  4. Logic check: Does the conclusion follow from the evidence? Are there unstated assumptions?
  5. Completeness check: Does the finding address the full question, or only part of it?
  6. Counter-evidence search: Actively search for information that contradicts the claim.

Verdict criteria

  • CONFIRMED: Independently verified via separate source(s). Evidence solid.
  • DISPUTED: Found counter-evidence, logical flaw, or source misrepresentation. Must provide specific counter-evidence.
  • UNVERIFIABLE: Cannot independently confirm or deny. Source inaccessible, claim too vague, or insufficient evidence either way.
  • OUTDATED: Claim was true but circumstances have changed. Provide current info.

Read the full file on GitHub · 101 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 · 101 lines · 37 tokens per session scan A e62485aca75f

Subscribe to this mod's changes

research-verifier is an agent published in the GitHub repository alphaelements/handoff-mcp (0 stars, last pushed 4d ago), licensed MIT. It adds 37 tokens to every session and 871 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.