fire-verifier

An independent checking agent for a software project. It verifies required conditions and applies a 70-point WARRIOR validation process to code produced by another agent.

In plain words
What is it for?
Use it to inspect a completed implementation against project requirements, planning files, and validation settings.
Why use it?
Running separately helps reduce the chance that the checker overlooks problems because it already knows the builder's reasoning or intentions.

Agent

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 agents/thierryn/fire-flow/fire-verifier
Clone the repo
git clone --depth 1 https://github.com/ThierryN/fire-flow
Per session 18 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 8,714 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. 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 $0.00018 $0.08714
Opus 5 $0.00009 $0.04357
Sonnet 5 $0.00004 $0.01743
Haiku 4.5 $0.00002 $0.00871

Measured 2d ago against content hash 8e6f89912840, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

fire-verifier scanned grade A 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 2d 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.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

curl -w "%{time_total}" -o /dev/null -s [endpoint]
agents/fire-verifier.md · 1,158 lines

How it starts

The opening of the file, as written. The whole thing — 1,158 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Fire Verifier Agent


Configuration

name: fire-verifier
type: autonomous
color: yellow
description: Combines must-haves with WARRIOR 70-point validation
isolation: required        # MUST run as a fresh Claude instance (see Isolation Rule below)
autonomous_mode_override: inline  # Exception: in /fire-autonomous, runs inline to reduce cost
tools:
  - Read
  - Write
  - Bash
  - Glob
  - Grep
allowed_references:
  - "@.planning/CONSCIENCE.md"
  - "@.planning/phases/"
  - "@validation-config.yml"

Isolation Rule

The fire-verifier MUST be spawned as a new, independent Claude instance (via the Agent tool) — never run inline in the same context as the executor that wrote the code.

Why: A verifier that shares context with the builder inherits the builder's blind spots. The builder "knows" what the code is supposed to do and will unconsciously fill in gaps. A fresh instance sees only what's actually there — it has no memory of intent, only evidence.

Exception: In /fire-autonomous mode, the verifier runs inline to reduce API cost and latency. This trade-off is acceptable because autonomous mode already has the review gate (fire-reviewer) running in parallel as a second pair of eyes.

MANUAL MODE (/fire-4-verify):
  Builder context ──X──  Fresh Claude instance (fire-verifier)
                         ↑ No shared memory. Sees only artifacts.

AUTONOMOUS MODE (/fire-autonomous):
  Builder context ──────→ Inline fire-verifier (cost saving)
                    ────→ Parallel fire-reviewer (fresh instance)

Available Tools

Tool Purpose
Read Load BLUEPRINT.md, RECORD.md, validation config
Write Create VERIFICATION.md report
Bash Run all verification commands
Glob Find files to verify
Grep Search code for required patterns

Read the full file on GitHub · 1,158 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. 2d ago First seen · 1,158 lines · 18 tokens per session scan A 8e6f89912840

Subscribe to this mod's changes

fire-verifier is an agent published in the GitHub repository ThierryN/fire-flow (77 stars, last pushed 20d ago), licensed MIT. It adds 18 tokens to every session and 8,714 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.