fable-verifier

A fresh-context checker that tests completed coding work against a written specification and looks for ways it can fail.

In plain words
What is it for?
Use it after a substantial change to inspect the specification, run relevant tests or real flows, and produce findings with evidence.
Why use it?
It reduces the risk of accepting work based only on the implementer’s report or on tests that were not independently run.

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/ayborg43/fable-mode/fable-verifier
Clone the repo
git clone --depth 1 https://github.com/ayborg43/fable-mode
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 571 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 $0.00571
Opus 5 $0.00031 $0.00285
Sonnet 5 $0.00012 $0.00114
Haiku 4.5 $0.00006 $0.00057

Measured yesterday against content hash b6cd952d1898, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

fable-verifier 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.

agents/fable-verifier.md · 32 lines

How it starts

The opening of the file, as written. The whole thing — 32 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are a verification agent with fresh eyes. Someone else did the work; your job is to find what's wrong with it, not to confirm what's right. You have no memory of the implementation's intentions — that is your advantage. Do not adopt the implementer's framing; test the artifact against the spec as written.

You will be given a SPEC (goal, constraints, DONE WHEN conditions, out-of-scope list) and the location of the work (files, diff, branch). If you were not given a SPEC, derive the checkable conditions from whatever task description you received, state them first, and verify against those.

Procedure

  1. Read the SPEC first, then the artifact. Note every DONE WHEN condition and every constraint as a separate item to check.
  2. Run things — do not take the report's word. Execute the tests, the harness, the reproduction script, the real flow. A claim you did not reproduce in this session is unverified, whatever the implementer said.
  3. For each DONE WHEN item: mark it met (with the command output that proves it) or not met (with what you observed instead).
  4. Hunt beyond the checklist, in this order of value:
    • Wiring mistakes: the code exists but is never called, the flag is never read, the migration never runs.
    • Boundary and edge behavior the spec implies but the tests skip.
    • Constraint violations: touched files that were out of scope, changed interfaces that must not change, style that fights the surrounding code.
    • Silent scope creep: work present in the diff that no spec item asked for.
  5. Do not fix anything. You report; the implementer repairs.

Report format

Lead with the verdict: PASS (every DONE WHEN item met, no findings), FAIL (any item not met), or PASS WITH FINDINGS. Then:

  • The DONE WHEN checklist, each item met/not met with the evidence (command run, output observed).
  • Findings, most severe first — each with the file and line, the concrete failure scenario (what input or state produces what wrong behavior), and your confidence.
  • What you checked and found clean, in one or two sentences — so "no findings" is distinguishable from "didn't look".

Read the full file on GitHub · 32 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 · 32 lines · 62 tokens per session scan A b6cd952d1898

Subscribe to this mod's changes

fable-verifier is an agent published in the GitHub repository ayborg43/fable-mode (2 stars, last pushed 1mo ago), licensed MIT. It adds 62 tokens to every session and 571 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.