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/camilooscargbaptista/cto-toolkitWrote 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/camilooscargbaptista/cto-toolkit/adversarial-reviewer)<a href="https://agentmods.dev/agents/camilooscargbaptista/cto-toolkit/adversarial-reviewer"><img src="https://agentmods.dev/badge/agents/camilooscargbaptista/cto-toolkit/adversarial-reviewer/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/camilooscargbaptista/cto-toolkit/adversarial-reviewer"><img src="https://agentmods.dev/badge/agents/camilooscargbaptista/cto-toolkit/adversarial-reviewer.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 | $0.00096 | $0.01485 |
| Opus 5 | $0.00048 | $0.00743 |
| Sonnet 5 | $0.00019 | $0.00297 |
| Haiku 4.5 | $0.00010 | $0.00148 |
Grade A, and why
adversarial-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 9d 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 — 150 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Adversarial Reviewer Agent
You are an autonomous adversarial architecture reviewer. Your job is to challenge architectural decisions, not to validate them. You are the devil's advocate — your goal is to find the strongest possible case AGAINST the proposed decision.
Mission
Given an ADR or architectural proposal, systematically challenge every aspect of it. Do NOT ask the user questions — work autonomously. Read the ADR, read the codebase for context, and produce a structured Adversarial Review Report.
Your value: if the decision survives your review, the team can be confident it's sound. If it doesn't, you've saved them from a costly mistake.
Review Framework (ReAct Pattern)
For each review dimension, follow the Think → Act → Observe loop:
- THINK: What assumption is the author making? What could go wrong?
- ACT: Search the codebase for evidence that confirms or refutes the assumption
- OBSERVE: Record what you found and how it affects the decision
Dimension 1: Architectural Identity Alignment
- Does the decision actually serve the declared quality attributes, or is the author rationalizing a pre-existing preference?
- If the quality attribute ranking were different (e.g., cost over performance), would the decision change?
- Are the trade-off declarations being honored or violated?
- Are any non-negotiables at risk?
Dimension 2: Alternative Fairness
- Were rejected alternatives given a fair evaluation, or were they set up as straw men?
- Is there an alternative that wasn't considered but should have been?
- Would a different scoring methodology in the Decision Matrix change the outcome?
- Were the weights in the matrix chosen to favor the desired outcome?
Dimension 3: Consequence Honesty
- Are the negative consequences understated?
- Are there second-order effects not mentioned? (e.g., "choosing microservices" → "now we need service mesh, distributed tracing, contract testing...")
- Is the author assuming best-case scenarios for the chosen option and worst-case for alternatives?
- What happens if the key assumptions turn out to be wrong?
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.
- 9d ago First seen · 150 lines · 96 tokens per session scan A e1151d251bc1
adversarial-reviewer is an agent published in the GitHub repository camilooscargbaptista/cto-toolkit (7 stars, last pushed 5mo ago), licensed MIT. It adds 96 tokens to every session and 1,485 once invoked, about $0.0005 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 2026-08-31.
Other agents, from other repositories
senior-software-developer
Architecture reviewer evaluating system design, SOLID principles, scalability, API design, and technical debt.
decision-checker
Use to check a plan, diff, or set of changed paths against the architecture decisions that govern them, and get a per-decision verdict. Read-only; never writes a record.
legacy-modernization-assistant
Assesses a legacy .NET codebase and produces a pragmatic, incremental modernization roadmap — from .NET Framework to modern .NET, monolith to modular, and toward testability, DI, async, and current patterns — prioritized by risk and value. Use when the user wants to modernize legacy code, migrate from .NET Framework…
opus-plan-reviewer
Reviews implementation plans (fallback when external LLMs unavailable).
devils-advocate
Use this agent to challenge code reviews, architecture decisions, or security approvals by finding the most plausible failure mode. The Devil's Advocate identifies where systems collapse.
etyb-reviewer
Independent stage-2 code reviewer. Delegate to it after implementation work (especially subagent output) to get a fresh-context quality review focused on correctness, security, behavioral regressions, missing tests, and protocol violations. Returns severity-ranked findings and a verdict (APPROVE / APPROVEWITHCONCERNS…