falsify

A command that challenges a plan, pitch, product requirements document, or other set of claims from several skeptical angles. It produces a pre-mortem and a report of risks that could invalidate the idea.

In plain words
What is it for?
Use it to attack a release plan, startup pitch, product requirements document, investment argument, or hypothesis and identify which risks could kill it.
Why use it?
It helps expose ignored evidence, weak customer demand, unrealistic schedules, easy-to-copy advantages, and distribution problems before time is spent building.

Command

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 commands/dormstern/forge/falsify
Clone the repo
git clone --depth 1 https://github.com/dormstern/forge
Per session 62 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 773 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 $0.00062 $0.00773
Opus 5 $0.00031 $0.00387
Sonnet 5 $0.00012 $0.00155
Haiku 4.5 $0.00006 $0.00077

Measured yesterday against content hash 438b8971fe61, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

falsify 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 yesterday.

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.

commands/falsify.md · 54 lines

How it starts

The opening of the file, as written. The whole thing — 54 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/forge:falsify — Standalone Devil's Advocate

Execution mode: direct. This command has explicit steps below — execute them in order. Do not enter plan mode, do not write a plan file, do not call ExitPlanMode. The framework already did the planning; you just run the steps.

You dispatch the devils-advocate agent against any user-supplied artifact: a release plan, a startup pitch, a PRD, a thesis section, an investor email — anything with claims that can be killed.

Behavior

  1. Receive the artifact. The user pastes text, points to a file, or says "falsify the current shape." If they don't provide one, ask: "What should I attack?" with three categories: a plan, a pitch, or a hypothesis.

  2. Print a pre-flight line before dispatching the agent (so the user knows what's happening + what success looks like):

    ⚡ Falsifying your plan across 8 attack vectors. This is uncomfortable on purpose —
       each ⚠ is a real risk, each ✗ is a kill. Aim to survive 5/8.
    
  3. Dispatch devils-advocate with the full attack framework:

    • Cherry-Picking: "Show me the evidence you ignored."
    • False Champion: "Your champion has no budget. Or worse, no problem."
    • Cope Moat: "Your moat is a feature, a UX preference, or a launch advantage."
    • TAM Mirage: "Your TAM is one bad SQL query. The real wedge is 1/100th."
    • Weak Wedge: "Your wedge could be a feature inside three incumbents within 90 days."
    • Timeline Cope: "What looks like 3 days is 3 weeks. What looks like 3 weeks is a quarter."
    • Replicability: "A funded competitor clones this in 12 months."
    • Distribution: "You have product. Where's the channel?" Plus a pre-mortem (≥5 scenarios) and an evidence-skew audit.
  4. Render the kill report in the screenshottable format (monospace, Unicode borders, ✓/⚠/✗ markers, confidence score per finding, Falsifiability Index at the bottom). One-line verdict per attack, plus a 48-hour cheap test for each KILL or WOUND.

  5. Persist the report if state files exist:

    • Founder mode → append to progress.md under "Recent DA Findings"
    • PM mode → append to progress.md and releases.json current_cycle if shaping is in progress

Read the full file on GitHub · 54 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. yesterday First seen · 54 lines · 62 tokens per session scan A 438b8971fe61

Subscribe to this mod's changes

falsify is a command published in the GitHub repository dormstern/forge (6 stars, last pushed 3mo ago), licensed MIT. It adds 62 tokens to every session and 773 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 2026-08-31.