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.
npx agentmods add skills/t1/tdder/prove-me-wrongnpx skills add t1/tdder --skill prove-me-wronggit clone --depth 1 https://github.com/t1/tdderWrote 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/skills/t1/tdder/prove-me-wrong)<a href="https://agentmods.dev/skills/t1/tdder/prove-me-wrong"><img src="https://agentmods.dev/badge/skills/t1/tdder/prove-me-wrong.svg" alt="Measured on agentmods" 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 | $0.00077 | $0.00440 |
| Opus 5 | $0.00039 | $0.00220 |
| Sonnet 5 | $0.00015 | $0.00088 |
| Haiku 4.5 | $0.00008 | $0.00044 |
Grade A, and why
prove-me-wrong 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.
What it actually says
Prove Me Wrong
Devil's advocate mode: find the strongest case against the user's position.
How to Run a Session
- Find the most threatening objection first — not the most obvious, the most damaging. A strong objection threatens the core of the position, not a detail.
- Open with a brief label that names the type of objection (e.g. Hidden assumption:, Scope mismatch:, Contradicts X:). The label is a summary of what you found — derive it from the objection, don't pick a category first and fit an objection to it.
- State the objection directly. No softening, no "that's a great point, but". Lead with the problem.
- Wait for the user to respond before surfacing the next objection.
- Concede honestly if the response genuinely addresses the objection. Don't retreat to a weaker version of the same argument.
- Repeat with the next strongest objection.
When to Stop
Stop when you've run out of real objections. Say so explicitly:
"I don't have a stronger objection than what we've already covered."
The user can say something like "keep going" to push further.
Never nitpick. If the next candidate is about wording, minor edge cases, or implementation detail rather than a genuine flaw in the reasoning — stop. Around five objections is typically the ceiling before quality degrades into noise.
The user can also exit the mode at any time by saying something like "ok, I'm convinced", "stop", or "let's move on".
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.
- 5d ago First seen · 43 lines · 77 tokens per session scan A dd12049f5420
prove-me-wrong is a skill published in the GitHub repository t1/tdder (14 stars, last pushed yesterday), licensed Apache-2.0. It adds 77 tokens to every session and 440 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.
Other skills, from other repositories
refactor
Automated iterative code refactoring with swarm-orchestrated specialist agents including deep codebase discovery, confidence-scored code review, and security analysis. Use this skill when the user wants to improve existing code quality, clean up messy code, restructure, simplify, reduce tech debt, or perform…
feature-dev
Guided feature development with swarm-orchestrated codebase exploration, multi-perspective architecture design, implementation, and quality review. Use this skill when the user wants to build a new feature, add new functionality, implement a capability, or create something that doesn't exist yet. Triggers on requests…
feature-dev
Guided feature development with swarm-orchestrated codebase exploration, multi-perspective architecture design, implementation, and quality review.
fp-error-handling
This skill should be used when the user asks about "functional error handling", mentions "Result type", "Either type", "Option type", "Maybe monad", discusses replacing try-catch with types, or wants to handle errors without exceptions.
fp-pure-functions
This skill should be used when the user asks about "pure functions", "side effects", mentions "immutability", "referential transparency", discusses hidden state changes, or wants to make functions more predictable and testable.
hexagonal-domain-isolation
This skill should be used when the user asks about "domain isolation", "anti-corruption layer", mentions "keeping domain pure", discusses preventing infrastructure leakage into business logic, or wants to protect the domain model from external concerns.