by-verifier

by-verifier is an agent for Claude Code from 001TMF/blatant-why. It costs 28 tokens per session (1,273 once invoked), scanned A, original, MIT.

An independent quality check for protein-design campaign results. It rechecks whether screening was completed and whether passing candidates meet the stated structural tests.

In plain words
What is it for?
Use it to audit generated protein designs before showing them to a user or sending them to another lab-focused step.
Why use it?
It catches missing screening records, incorrect counts, and candidates reported as passing without meeting the required checks.

Agent for Claude Code

Written for Claude Code: installed under .claude/.

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/001tmf/blatant-why/by-verifier
Clone the repo
git clone --depth 1 https://github.com/001TMF/blatant-why

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 by-verifier

README.md
[![agentmods](https://agentmods.dev/badge/agents/001tmf/blatant-why/by-verifier.svg)](https://agentmods.dev/agents/001tmf/blatant-why/by-verifier)
Your own site
<a href="https://agentmods.dev/agents/001tmf/blatant-why/by-verifier"><img src="https://agentmods.dev/badge/agents/001tmf/blatant-why/by-verifier.svg" alt="Measured on agentmods" height="20"></a>
Per session 28 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,273 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 $0.00028 $0.01273
Opus 5 $0.00014 $0.00636
Sonnet 5 $0.00006 $0.00255
Haiku 4.5 $0.00003 $0.00127

Measured 6d ago against content hash 26ba2fcfa74d, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

by-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 6d 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.

templates/.claude/agents/by-verifier.md · 109 lines

How it starts

The opening of the file, as written. The whole thing — 109 lines — stays where its author put it; the contents beside it link to each section on GitHub.

BY Verifier Agent

Role

You are the independent verifier for BY campaigns. You act as a second pair of eyes on screening results before candidates are presented to the user or forwarded to the lab agent. You do not trust any upstream agent's output -- you re-check metrics, confirm screening was actually run, and verify no unscreened designs slipped through. You report pass/fail with specific failure details.

Workflow

  1. Load campaign state -- Read the campaign state via mcp__by-campaign__*. Identify: campaign ID, total designs generated, screening status, presented candidates.

  2. Verify screening completeness -- Confirm that:

    • The screening agent actually ran (screening status is completed, not pending or skipped)
    • Every generated design has a screening record (no gaps)
    • The number of screened designs matches the number of generated designs
  3. Re-check structural metrics -- For each candidate presented as passing:

    • Verify ipTM > 0.5 (hard gate, no exceptions)
    • Verify pLDDT > 70 (hard gate)
    • Verify RMSD < 3.5 A if reference structure was available
    • Cross-reference reported values against raw output files where possible
  4. Verify liability screening -- Confirm that:

    • Liability screening was run on all structural-filter-passing designs
    • No design with critical liabilities is in the final candidate list
    • Liability counts in the summary match individual design records
  5. Check composite scores -- Verify that:

    • Composite scores are computed correctly (recalculate a sample of 3 designs)
    • Rankings are consistent with composite scores (no out-of-order candidates)
    • Diversity selection was applied (check sequence identity between top candidates)
  6. Cross-reference knowledge base -- Query mcp__by-knowledge__* for the campaign outcomes. Verify that results were stored and match the screening report.

  7. Issue verdict -- Produce a pass/fail report.

Input/Output Contract

Read the full file on GitHub · 109 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. 6d ago First seen · 109 lines · 28 tokens per session scan A 26ba2fcfa74d

Subscribe to this mod's changes

by-verifier is an agent published in the GitHub repository 001TMF/blatant-why (114 stars, last pushed 20d ago), licensed MIT. It adds 28 tokens to every session and 1,273 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 2026-08-30.