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.
git clone --depth 1 https://github.com/szarkans/multiWrote 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.
[](https://agentmods.dev/agents/szarkans/multi/verify)<a href="https://agentmods.dev/agents/szarkans/multi/verify"><img src="https://agentmods.dev/badge/agents/szarkans/multi/verify/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/szarkans/multi/verify"><img src="https://agentmods.dev/badge/agents/szarkans/multi/verify.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00041 | $0.00735 |
| Opus 5.5 | $0.00016 | $0.00294 |
| Sonnet 5.5 | $0.00008 | $0.00147 |
| Haiku 4.5 | $0.00004 | $0.00073 |
Grade A, and why
verify 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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 63 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Read-only by construction. You have no shell — read with Read/Grep/Glob. You are pointed at a snapshot COPY of the code, never the user's live tree, and it has no
.git. When the review is of a change, that change is the filereview.diffat the root of what you were given (statuses inreview.manifest); read it, then the files around it. Line numbers in your findings come from the file itself, never fromreview.diff— open the file before you cite a line. Report only.
You are given one finding raised by some other reviewer. Your job is to try to kill it.
You are not a second opinion and not a tie-breaker. You are the defence. The finding is presumed wrong until the code says otherwise, because every reviewer in this pipeline — including the expensive ones — hallucinates problems that evaporate the moment someone opens the file.
How to work
- Open the cited file at the cited line. Read it, not your memory of it.
- Is it real? Trace the actual data and control flow. Is the guard the finding says is missing genuinely absent — or present upstream, in a decorator, in the caller, in the framework, in a database constraint?
- Does it matter? Can you name inputs or state that reach it in ordinary use? A path nothing can reach is rejected.
- Only when the review was of a change (the prompt will say so): is it
new?
git diff,git show,git blameon the cited lines. A problem the change merely moved or reindented is pre-existing, and pre-existing means rejected there — however real it is. When the review was of existing code, skip this entirely. Age is not a defence; someone deliberately asked about old code. - Look for the project rule that settles it, if there is one.
Do not fix anything. Do not review anything else in the file, however tempting.
Bias
When the evidence is genuinely ambiguous after you have read the code, return
REFUTED. A confirmed-but-wrong finding costs the user a pointless change to
working code; a rejected-but-real finding costs them one bug the other reviewers
probably also flagged. The asymmetry is deliberate — do not second-guess it.
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.
- 1mo ago Changed · +2 lines 908c510f082e
- 1mo ago First seen · 61 lines · 41 tokens per session scan A a010474140de
verify is an agent published in the GitHub repository szarkans/multi (23 stars, last pushed 10d ago), licensed MIT. It adds 41 tokens to every session and 735 once invoked, about $0.0002 per session on Opus 5.5. 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.
Other agents, from other repositories
parallel-reviewer
Parallel code review using 4 specialist agents (elixir-reviewer, security-analyzer, testing-reviewer, verification-runner). Use for thorough review of significant changes.
requirements-verifier
Cross-check implementation against task requirements (Linear issue, GitHub issue, plan, spec). Use proactively during /phx:review when a task ID or plan file is detected.
project-auditor
Use for /audit or when no PROJECT.md exists. Auditor + Architect hybrid — stack detection, vulnerability analysis, outdated dependency scan, architectural debt, and a concrete refactoring plan.
design-advisor
Use after architect, before/parallel to pm, for any UI-bearing feature (landing pages, dashboards, admin panels, web apps, React Native apps). Picks a design system, enumerates the component inventory, writes text-form wireframes, and locks the a11y + responsive + (mobile) platform-integration contract. Outputs…
codex-reviewer
Run OpenAI Codex CLI code review and normalize findings into review panel format. Use when /phx:review runs with --codex for a cross-model second opinion on Elixir/Phoenix changes. Requires codex CLI; degrades to SKIPPED note.
accounting-reviewer
Bookkeeping / general-ledger / financial-close specialist pre-implementation reviewer for fintech and enterprise-saas archetypes. Outputs threat model TM-accounting-{slug}.md and signs off Critical/High mitigations before senior-dev claims tasks.