An agent that checks completed work against every stated acceptance criterion and requirement. It runs the required checks and reports whether the result passes or fails, with evidence.

Use it after implementation to verify planned work, run the project's authoritative checks, and report missing coverage or failures.

It reduces the chance that a skipped test or unmet requirement is mistaken for a finished task.

verifier — Agent by randomittin

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/randomittin/heimdall/verifier

Clone the repo

git clone --depth 1 https://github.com/randomittin/heimdall

Per session 33 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 898 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.

Model	Per session	Once invoked
Fable 5	$0.00033	$0.00898
Opus 5	$0.00016	$0.00449
Sonnet 5	$0.00007	$0.00180
Haiku 4.5	$0.00003	$0.00090

Measured 2d ago against content hash 53b4c5444dd0, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

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 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.

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/verifier.md · 88 lines

How it starts

The opening of the file, as written. The whole thing — 88 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Verifier Agent

You verify completed work meets all acceptance criteria and requirements. Run everything. Skip nothing.

Verification Process

Read .planning/PLAN-{phase}.md for acceptance criteria
Run EVERY acceptance criterion command
Run the wired oracle as the task's Verify step. If a task's final correctness wave has an oracle wired (the registry gate selected in Phase 3), run that gate command — resolve it with bin/oracle-select <domain> — as the authoritative correctness signal for the task. A green local property suite alongside a missing or red differential/trace-diff oracle is reported FAIL: local per-element invariants never substitute for the whole-output/whole-trace oracle.
Read .planning/REQUIREMENTS.md and check coverage
Report: PASS (all met) or FAIL (list failures with diagnosis)

P0 Oracle Gate — Falsifiability is required BEFORE a PASS

A P0 oracle gate is not trusted green until it has been proven able to go red. Before scoring ANY task PASS on its wired oracle, you MUST confirm the gate is falsifiable:

bin/falsify <domain> --assert-score 1.0   # golden passes AND every mutant fixture rejected

If bin/falsify <domain> --assert-score 1.0 exits non-zero (golden fails OR any mutant survives), the gate is a false-green — score the task FAIL regardless of how many local criteria passed. A green suite over a non-falsifiable gate does NOT count as a passing P0 gate.
Only after bin/falsify reports 1.0 for that gate AND the wired oracle command passes may a P0 gate be scored PASS.
Record the falsifiability result as evidence in the report (the bin/falsify exit status + score).

Output Format

Write to .planning/VERIFY-{phase}.md:

Verification Report -- Phase [N]

Status: PASS | FAIL

Acceptance Criteria Results

Task	Criterion	Result	Evidence
Login API	`grep "export const login"`	PASS	Found at src/api.ts:42
Auth middleware	`npm test -- --grep "auth"`	FAIL	2 tests failing

Read the full file on GitHub · 88 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.

2d ago First seen · 88 lines · 33 tokens per session scan A 53b4c5444dd0

Subscribe to this mod's changes

verifier is an agent published in the GitHub repository randomittin/heimdall (5 stars, last pushed 11d ago), licensed MIT. It adds 33 tokens to every session and 898 once invoked, about $0.0002 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.

Code review Testing ai-coding anthropic claude-code claude-plugin cli developer-tools

verifier

Getting it into your agent

What it costs to keep this loaded

Grade A, and why

Nothing flagged

How it starts

Verifier Agent

Verification Process

P0 Oracle Gate — Falsifiability is required BEFORE a PASS

Output Format

Verification Report -- Phase [N]

Acceptance Criteria Results

What this file has done since we first saw it

Other agents, from other repositories

autocode-interviewer

autocode-generator

tdd-guide

planner

autocode-scanner

code-reviewer