fact-gate

fact-gate is a skill for Claude Code, Codex from chadcorp/warden. It costs 0 tokens per session (504 once invoked), scanned A, original, Apache-2.0.

A checking process for confirming factual, legal, and financial claims against dated primary sources before publication.

In plain words
What is it for?
Use it to extract claims from a draft, classify them, and verify each one against an authoritative source.
Why use it?
It catches unsupported or outdated numbers, dates, rates, rules, deadlines, and jurisdiction details before they reach readers.

Skill for Claude CodeCodex

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 skills/chadcorp/warden/fact-gate
Any agent
npx skills add chadcorp/warden --skill fact-gate
Clone the repo
git clone --depth 1 https://github.com/chadcorp/warden

Made for: Claude Code, Codex.

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 fact-gate

README.md
[![agentmods](https://agentmods.dev/badge/skills/chadcorp/warden/fact-gate.svg)](https://agentmods.dev/skills/chadcorp/warden/fact-gate)
Your own site
<a href="https://agentmods.dev/skills/chadcorp/warden/fact-gate"><img src="https://agentmods.dev/badge/skills/chadcorp/warden/fact-gate.svg" alt="Measured on agentmods" height="20"></a>
Per session 0 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 504 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 5.1 $0.00000 $0.00504
Opus 5 $0.00000 $0.00252
Sonnet 5 $0.00000 $0.00101
Haiku 4.5 $0.00000 $0.00050

Measured 4d ago against content hash 86213dfce1ff, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-05, from the pricing page.

Security

Grade A, and why

fact-gate 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 4d 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.

skills/research-brain/fact-gate/SKILL.md · 47 lines

How it starts

The opening of the file, as written. The whole thing — 47 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Fact Gate

Verify every factual, legal, or financial claim against a dated primary source before anything ships. This is a blocking discipline: a draft does not pass until each checkable claim is signed off.

When to use

Any deliverable that cites laws, official figures, statistics, regulations, rates, or deadlines — or whenever someone asks "is this accurate?" / "check this against sources."

The protocol

  1. Extract the checkable claims. Walk the draft and pull out every assertion a reader could dispute: numbers, dates, rates, thresholds, named rules, jurisdictions. Restate each as a single, testable sentence.

  2. Class each claim. Tag the kind of claim, because each kind has a typical failure mode:

    • Rate / amount — often stale; verify the year it applies to.
    • Threshold / phase-out limit — easy to off-by-one; confirm the exact band.
    • Deadline — confirm the current cycle, not last year's.
    • Jurisdiction / forum — confirm the level (federal vs state, court vs agency); loose phrasing from a single outlet is a classic trap.
  3. Match each claim to a dated primary source. Prefer the authoritative issuer (the statute, the official table, the regulator's own page) over a secondary summary. Record the source and its date next to the claim.

  4. Block on the unverifiable. If a claim cannot be tied to a primary source, it does not ship as fact. Either soften it to attributed opinion ("according to X…"), or cut it. Never assert a number you could not source.

  5. Re-read for the silent claim-classes. Court forum, effective dates, "up to" limits, and percentages are the ones a casual pass skips. Sweep for them explicitly.

Output

A claims ledger: each claim, its class, its dated source, and a status of SIGNED / SOFTENED / CUT. The draft passes only when no claim is left unsigned. If anything is post-dated beyond your knowledge or unsourced, say so plainly rather than guessing.

Honesty rule

"I could not verify this" is a valid, valuable result. A fact gate that waves everything through is not a gate.

Read the full file on GitHub · 47 lines

Files

What ships with it

3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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. 4d ago First seen · 47 lines · 0 tokens per session scan A 86213dfce1ff

Subscribe to this mod's changes

fact-gate is a skill published in the GitHub repository chadcorp/warden (0 stars, last pushed 2mo ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 504 tokens. 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.