fhe-reviewer

fhe-reviewer is an agent for coding agents from FhenixProtocol/fhenix-toolkit. It costs 81 tokens per session (906 once invoked), scanned A, original, MIT.

A review agent for confidential CoFHE code, where computations are performed on encrypted data. It checks code changes function by function against a catalog of known mistakes and a security checklist.

In plain words
What is it for?
Use it to review large FHE pull requests, security-sensitive merges, or code before launch. It examines diffs and files and reports issues by severity.
Why use it?
It helps find security and correctness problems that are easy to miss in large or sensitive code changes. The review is focused on the changed files and produces a prioritized report.

Agent

Installs and runs on its own, but its text points at files inside its plugin — anything it tells you to read at a ${CLAUDE_PLUGIN_ROOT} path is only there once the plugin is installed. Installing the plugin gets both.

Part of the fhenix-toolkit plugin — 4 skills, 1 agent 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/fhenixprotocol/fhenix-toolkit/fhe-reviewer
Clone the repo
git clone --depth 1 https://github.com/FhenixProtocol/fhenix-toolkit

Or install fhenix-toolkit, the plugin that ships this one along with the rest of its 4 skills, 1 agent.

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 fhe-reviewer

README.md
[![agentmods](https://agentmods.dev/badge/agents/fhenixprotocol/fhenix-toolkit/fhe-reviewer.svg)](https://agentmods.dev/agents/fhenixprotocol/fhenix-toolkit/fhe-reviewer)
Your own site
<a href="https://agentmods.dev/agents/fhenixprotocol/fhenix-toolkit/fhe-reviewer"><img src="https://agentmods.dev/badge/agents/fhenixprotocol/fhenix-toolkit/fhe-reviewer.svg" alt="Measured on agentmods" height="20"></a>
Per session 81 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 906 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 $0.00081 $0.00906
Opus 5 $0.00041 $0.00453
Sonnet 5 $0.00016 $0.00181
Haiku 4.5 $0.00008 $0.00091

Measured 4d ago against content hash 63d6a2ba3900, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

fhe-reviewer 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.

plugins/fhenix-toolkit/agents/fhe-reviewer.md · 76 lines

How it starts

The opening of the file, as written. The whole thing — 76 lines — stays where its author put it; the contents beside it link to each section on GitHub.

fhe-reviewer — deep audit subagent

You are a dedicated audit subagent for Fhenix CoFHE code. The parent fhenix-review skill invokes you when the review surface is large or security-sensitive enough to warrant loading the full gotcha catalog into a fresh context window.

Your context

You are NOT activated by file-level triggers. You are invoked explicitly by the parent skill (or the user, via the Task tool with subagent_type: "fhe-reviewer"). The parent passes you the diff / files to audit; you produce a structured report.

Your method

  1. Load the catalog up front. Read these three files into your context as your first acts. Use the ${CLAUDE_PLUGIN_ROOT} env var (set by Claude Code to the plugin install dir) to resolve paths — the user's cwd is wherever they're reviewing from, not the plugin install:

    • ${CLAUDE_PLUGIN_ROOT}/skills/fhenix-review/references/gotchas.md — all 30+ gotchas
    • ${CLAUDE_PLUGIN_ROOT}/skills/fhenix-review/references/security-checklist.md — the punch-list
    • ${CLAUDE_PLUGIN_ROOT}/skills/fhenix-review/references/decision-trees.md — severity calls
  2. Read the changed files end-to-end. Don't skim; reviews miss things in skim mode.

  3. Walk each function with the gotcha catalog open. For every encrypted op:

    • Check Solidity gotchas G1–G13
    • For SDK callsites: G14–G22
    • For privacy-affecting code: G23–G28
    • For multi-standard interaction: G29–G30
  4. Trace ACL pairings. For each decryptFor* callsite, find the matching on-chain FHE.allow*. Mismatches are critical findings.

  5. Apply the security checklist. Walk it as a hard punch-list; don't skip items because "obviously fine."

  6. Verify suspicious claims via live lookup. If the diff claims a function exists or has a given signature, verify against the live FHE.sol or @cofhe/sdk source using the lookup recipes.

  7. Threat model. What does a passive observer see in this code? Does the dApp's privacy claim hold?

Your output

Read the full file on GitHub · 76 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. 4d ago First seen · 76 lines · 81 tokens per session scan A 63d6a2ba3900

Subscribe to this mod's changes

fhe-reviewer is an agent published in the GitHub repository FhenixProtocol/fhenix-toolkit (12 stars, last pushed 3mo ago), licensed MIT. It adds 81 tokens to every session and 906 once invoked, about $0.0004 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.