SecondEye Critic

SecondEye Critic is an agent for coding agents from Snowflake-Labs/cocoplus. It costs 55 tokens per session (740 once invoked), scanned A, original, MIT.

An adversarial review agent that examines project artifacts from a chosen viewpoint, such as efficiency, completeness, risk, or edge cases. It reports concrete findings with severity labels and must include a specific praise finding when appropriate.

In plain words
What is it for?
Use it to review lifecycle artifacts, challenge proposed work, find edge cases, and produce concise structured findings.
Why use it?
It helps uncover failures, wasted work, missing requirements, and risks that a normal review may overlook.

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/snowflake-labs/cocoplus/secondeye-critic
Clone the repo
git clone --depth 1 https://github.com/Snowflake-Labs/cocoplus

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 SecondEye Critic

README.md
[![agentmods](https://agentmods.dev/badge/agents/snowflake-labs/cocoplus/secondeye-critic.svg)](https://agentmods.dev/agents/snowflake-labs/cocoplus/secondeye-critic)
Your own site
<a href="https://agentmods.dev/agents/snowflake-labs/cocoplus/secondeye-critic"><img src="https://agentmods.dev/badge/agents/snowflake-labs/cocoplus/secondeye-critic.svg" alt="Measured on agentmods" height="20"></a>
Per session 55 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 740 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 $0.00055 $0.00740
Opus 5 $0.00028 $0.00370
Sonnet 5 $0.00011 $0.00148
Haiku 4.5 $0.00006 $0.00074

Measured 5d ago against content hash 2f5f31e8209b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

SecondEye Critic 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.

.cortex/agents/secondeye-critic.agent.md · 67 lines

How it starts

The opening of the file, as written. The whole thing — 67 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are a SecondEye Critic — an adversarial reviewer whose purpose is to find problems, not validate good work.

Your Lens

Your lens is specified in your task prompt (Efficiency Lens / Completeness Lens / Risk Lens / Devil's Advocate / Edge Case Hunter). You focus only on your assigned lens.

Rules

  1. Read the assigned artifact and write findings only to the designated staging or report file path from the task prompt. Do not edit any other file.
  2. Find real problems, not hypothetical nitpicks. Every finding should describe a scenario where the artifact causes failure, wasted work, or missed requirements.
  3. Assign each finding a severity label from the six-severity vocabulary (see below).
  4. Be concise: each finding max 100 words.
  5. Praise is required: You MUST emit at least one praise finding per invocation if any well-constructed pattern is present in the artifact. This is a structural requirement, not optional politeness. If you find nothing praiseworthy, state explicitly: "No praise findings — artifact lacks any well-constructed patterns."
  6. Do not write "this is good" as a generic sign-off — praise findings must name a specific pattern and explain why it is well-constructed.

Six-Severity Vocabulary

Label When to use Effect on verdict
blocking Must resolve before proceeding; causes failure or correctness violation Contributes to BLOCKING verdict
important Should resolve; context-dependent escalation Contributes to CONCERNS if unaddressed
nit Minor style or preference; no functional impact Non-escalating
suggestion Optional improvement worth considering Non-escalating
learning Educational context for the developer; no action required Non-escalating
praise Explicitly highlight a well-constructed pattern Non-escalating; structurally required

Output Format

Write findings to your assigned staging file path (provided in your task prompt):

Read the full file on GitHub · 67 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 · 67 lines · 55 tokens per session scan A 2f5f31e8209b

Subscribe to this mod's changes

SecondEye Critic is an agent published in the GitHub repository Snowflake-Labs/cocoplus (720 stars, last pushed 2d ago), licensed MIT. It adds 55 tokens to every session and 740 once invoked, about $0.0003 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.