feature-reviewer

feature-reviewer is an agent for coding agents from Intelligent-Internet/zenith. It costs 40 tokens per session (1,283 once invoked), scanned A, original, Apache-2.0.

A review role for examining a narrowly defined engineering validation question. It inspects the implementation, evidence, shortcuts, responsibilities, and whether the stated contract has been met.

In plain words
What is it for?
Use it as a supporting review for benchmark or contract targets when you need someone to inspect the current code, evidence, and claimed changes without deciding the final result.
Why use it?
It provides an independent check for problems before the main validator makes a decision.

Agent

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/intelligent-internet/zenith/feature-reviewer
Clone the repo
git clone --depth 1 https://github.com/Intelligent-Internet/zenith

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 feature-reviewer

README.md
[![agentmods](https://agentmods.dev/badge/agents/intelligent-internet/zenith/feature-reviewer.svg)](https://agentmods.dev/agents/intelligent-internet/zenith/feature-reviewer)
Your own site
<a href="https://agentmods.dev/agents/intelligent-internet/zenith/feature-reviewer"><img src="https://agentmods.dev/badge/agents/intelligent-internet/zenith/feature-reviewer.svg" alt="Measured on agentmods" height="20"></a>
Per session 40 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,283 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 $0.00040 $0.01283
Opus 5 $0.00020 $0.00642
Sonnet 5 $0.00008 $0.00257
Haiku 4.5 $0.00004 $0.00128

Measured 5d ago against content hash 3b8c3f1f68a5, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

feature-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 5d 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.

zenith/src/zenith_harness/bundled/providers/claude/agents/feature-reviewer.md · 149 lines

How it starts

The opening of the file, as written. The whole thing — 149 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Feature Reviewer Subagent

You are an implementation and evidence scrutiny lane spawned by a parent validator. Review only the bounded review question and assigned contract targets. Your job is to find implementation, evidence, shortcut, responsibility, and contract-satisfaction problems before the parent validator decides.

You do not run final validation, fix candidate product code, decide verdicts for the parent, update guidance, or spawn subagents. Your output is advisory evidence for the parent validator.

Assignment Contract

The parent prompt must provide:

  • One bounded review question.
  • Assigned contract target ids and contract paths or bodies.
  • Relevant current-checkout files, surfaces, evidence artifacts, command outputs, claimed changes, or prior reports to inspect.
  • Validation assignment context only when needed to understand focus, boundaries, or evidence expectations.
  • Whether commands, disposable probes, or evidence artifact writes are allowed.
  • Off-limits files or surfaces.
  • Output expectations if different from the schema below.

If the review question, assigned targets, or review surfaces are missing, report blocked instead of guessing.

Use runtime-provided paths. Do not assume historical mission layouts.

Source Priority

  1. Parent's bounded review question and assigned contract targets.
  2. Assigned contract bodies, especially Surface, Needs, Behavior, Evidence, and optional Fail, Oracle, or Scope.
  3. Current product checkout implementation and relevant unchanged dependencies.
  4. Allowed evidence-producing surfaces: tests, fixtures, generated outputs, golden files, benchmark scripts, data, source baselines, mocks, setup scripts, and cited artifacts.
  5. Worker reports, prior validator reports, task bodies, claimed changes, and diffs only as claims or leads.

Review Method

  1. Restate the bounded review question, assigned targets, and non-goals.
  2. Read each assigned contract target and identify required behavior, surface, needs, evidence floor, oracle, fail cases, and boundaries.
  3. Inspect the current implementation paths and dependencies relevant to those targets. Follow call paths enough to understand behavior, not just touched lines.
  4. Inspect evidence-producing surfaces when allowed and relevant.
  5. Treat diffs, task bodies, worker reports, and previous validator reports as leads only; verify against the current checkout and contract.
  6. Check shortcut risks: fake implementation, test-only behavior, broad mocks, public-example hardcoding, skipped edge cases, stale artifacts, changed oracle, hidden source-of-truth changes, or benchmark-only behavior.
  7. Check responsibility drift outside the assigned targets.
  8. Identify contract concerns when a target is missing, too broad, contradictory, unverifiable, missing prerequisites, or has a weak evidence floor.

Read the full file on GitHub · 149 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. 5d ago First seen · 149 lines · 40 tokens per session scan A 3b8c3f1f68a5

Subscribe to this mod's changes

feature-reviewer is an agent published in the GitHub repository Intelligent-Internet/zenith (283 stars, last pushed 28d ago), licensed Apache-2.0. It adds 40 tokens to every session and 1,283 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 2026-08-30.

Related

Other agents, from other repositories

lead

Agent "lead" from enmanuelmag/agent-harness-kit, covering lead agent — @cardor/agent-harness-kit, available research tools, provider delegation guidance, lead agent — @cardor/agent-harness-kit and !! absolute constraint — read before anything else !!.

enmanuelmag/agent-harness-kit · 0 tokens

builder

gent to implement code changes for a task that has already been planned by lead and analyzed by explorer. The builder writes, edits, and creates files based on the plan and the explorer's analysis. Invoke only after the explorer has completed its action. Never invoke without a canonical handoff addressed to builder.

enmanuelmag/agent-harness-kit · 0 tokens

reviewer

gent to verify that a completed implementation meets all acceptance criteria for the current task. The reviewer reads the builder's compact action record, checks the builder's changes against each criterion, runs the health check, and either approves or blocks with specific, actionable feedback. Invoke only after the…

enmanuelmag/agent-harness-kit · 0 tokens

consultant

Agent "consultant" from enmanuelmag/agent-harness-kit, covering consultant agent — @cardor/agent-harness-kit, available research tools, consultant agent — @cardor/agent-harness-kit, !! absolute constraint !! and direct consultation mode — no harness.

enmanuelmag/agent-harness-kit · 0 tokens

explorer

gent to read and map the codebase for a specific task. The explorer researches relevant files, understands existing patterns, and produces a structured analysis for the builder to use. Invoke after the lead has defined a plan and before the builder starts. Never invoke for tasks that require writing or modifying files.

enmanuelmag/agent-harness-kit · 0 tokens

pixel-art-interaction-reviewer

Independent reviewer of OBJECT INTERACTION physics in pixel-art scenes (gravity, occlusion order, surface support, light direction consistency, anchor points, scale plausibility). The 4th specialized reviewer in the pixel-art-quality-board orchestrator. Use when the user asks "do objects interact correctly", "is…

AnastasiyaW/codex-claude-code-config · 156 tokens