devils-advocate

devils-advocate is an agent for Claude Code from VandanaAjayDubey111/great-pm. It costs 159 tokens per session (2,503 once invoked), scanned B, original, MIT.

An adversarial questioner that looks for hidden assumptions in a product idea or decision. It challenges the framing of the problem without judging the person who proposed it.

In plain words
What is it for?
Use it to test the foundations of an idea, plan, or proposed direction. It produces a ranked set of questions instead of approving, rejecting, or rechecking the finished work.
Why use it?
It helps expose important questions that were missed before a decision is made, especially questions whose answers could change the decision.

Agent for Claude Code

Written for Claude Code: effort in frontmatter. Also seen: model in frontmatter.

Part of the great-pm plugin — 10 commands, 48 agents shipped together

Good fit Use it to test the foundations of an idea, plan, or proposed direction. It produces a ranked set of questions instead of approving, rejecting, or rechecking the finished work.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/vandanaajaydubey111/great-pm/devils-advocate
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.

Clone the repo
git clone --depth 1 https://github.com/VandanaAjayDubey111/great-pm

Made for: Claude Code.

Or install great-pm, the plugin that ships this one along with the rest of its 10 commands, 48 agents.

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 devils-advocate

README.md
[![agentmods](https://agentmods.dev/badge/agents/vandanaajaydubey111/great-pm/devils-advocate/github.svg)](https://agentmods.dev/agents/vandanaajaydubey111/great-pm/devils-advocate)
Your own site
<a href="https://agentmods.dev/agents/vandanaajaydubey111/great-pm/devils-advocate"><img src="https://agentmods.dev/badge/agents/vandanaajaydubey111/great-pm/devils-advocate/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.

agentmods 80×15 button for devils-advocate

Your own site · 80×15
<a href="https://agentmods.dev/agents/vandanaajaydubey111/great-pm/devils-advocate"><img src="https://agentmods.dev/badge/agents/vandanaajaydubey111/great-pm/devils-advocate.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 159 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,503 The whole file, excluding the scripts and references it only reads on demand.
Security scan B 1 finding. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00159 $0.02503
Opus 5 $0.00079 $0.01252
Sonnet 5 $0.00032 $0.00501
Haiku 4.5 $0.00016 $0.00250

Measured 10d ago against content hash 0054b7583d3e, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade B, and why

devils-advocate scanned grade B with 1 finding 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 10d 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.

Strips warnings and disclaimersmediumAnti-refusal

Omitting safety caveats hides risk from the user and is a common jailbreak preamble.

3. **File, don't lecture.** Each question that would BLOCK a sound decision gets
agents/devils-advocate.md · 184 lines

How it starts

The opening of the file, as written. The whole thing — 184 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are devils-advocate — great-pm's adversarial interrogator. Your job is to find the questions nobody asked: the assumptions an output rests on, the premises the human took for granted in how they framed the problem, the angles that went untested. You attack the IDEA's hidden foundations — never the person.

What makes you DIFFERENT (stay in your lane — this is what prevents a mess)

  • pm-reviewer reviews a FINISHED package for flaws and proposes fixes (reactive, on the artefact).
  • pm-advisor gives ONE external opinion on the bet (reactive, on the idea).
  • You go UPSTREAM of both. You interrogate the assumptions and the framing — including the human's — to surface what was never questioned. You do NOT re-review artefact quality (pm-reviewer's job) and do NOT render an opinion/verdict on whether the idea is good (pm-advisor's job). You produce QUESTIONS, ranked by whether their answer would flip the decision.

Governance (MANDATORY — overrides everything below)

You DRAFT and PROPOSE. You ask; you never decide, approve, or block on your own authority. Your questions inform the human and pm-lead. The skill-swap carve-out belongs to skill-scout, not to you.

The bounds that keep you an asset, not noise (NON-NEGOTIABLE)

  1. Every question must be able to CHANGE a decision. If a question's answer wouldn't alter what we do next, it does not make your list. No questioning for the sake of questioning. No trivia.
  2. CONVERGE — you are finite BY DEFAULT. Scan wide internally, but surface AT MOST 5–7 questions per round, ranked. Report the width of the scan ("considered N candidates across M angles — surfacing the 7 that flip the decision"), then state explicitly what SURVIVED ("the rest holds up"). You are not a spiral; you end. The human can widen you — see ATTACK HARDER below — but you never widen yourself.
  3. File, don't lecture. Each question that would BLOCK a sound decision gets filed into the Open-Decision Register (a beads open-decision), so it is tracked, not buried. Advisory ones are listed but not filed as blockers.
  4. Socratic, not hostile. Target the assumption, never the person. "What would have to be true for this to hold?" not "this is wrong."

Read the full file on GitHub · 184 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. 10d ago First seen · 184 lines · 159 tokens per session scan B 0054b7583d3e

Subscribe to this mod's changes

devils-advocate is an agent published in the GitHub repository VandanaAjayDubey111/great-pm (3 stars, last pushed 1mo ago), licensed MIT. It adds 159 tokens to every session and 2,503 once invoked, about $0.0008 per session on Opus 5. A static security scan graded it B with 1 finding (strips warnings and disclaimers). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.