implementation-verifier

implementation-verifier is an agent for coding agents from robisson/build-like-amazon-agent-skills. It costs 0 tokens per session (1,555 once invoked), scanned A, original, MIT.

A software verification reviewer that checks whether an implementation satisfies its written requirements. It uses ordinary tests, tests across many generated inputs, interface checks, and regression checks, rather than relying only on compilation or a few examples.

In plain words
What is it for?
Use it after implementation to derive testable rules from requirements and verify them. It helps test valid and invalid inputs, required responses, interface compatibility, and protection against regressions.
Why use it?
It catches cases where code appears to work for sample inputs but fails broader rules or breaks existing behavior. It turns requirements into checks that can reveal missing or incorrect behavior.

Agent

Part of the build-like-amazon plugin — 28 skills, 14 commands, 10 agents shipped together

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/robisson/build-like-amazon-agent-skills/implementation-verifier
Clone the repo
git clone --depth 1 https://github.com/robisson/build-like-amazon-agent-skills

Or install build-like-amazon, the plugin that ships this one along with the rest of its 28 skills, 14 commands, 10 agents.

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 implementation-verifier

README.md
[![agentmods](https://agentmods.dev/badge/agents/robisson/build-like-amazon-agent-skills/implementation-verifier.svg)](https://agentmods.dev/agents/robisson/build-like-amazon-agent-skills/implementation-verifier)
Your own site
<a href="https://agentmods.dev/agents/robisson/build-like-amazon-agent-skills/implementation-verifier"><img src="https://agentmods.dev/badge/agents/robisson/build-like-amazon-agent-skills/implementation-verifier.svg" alt="Measured on agentmods" height="20"></a>
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,555 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 $0.00000 $0.01555
Opus 5 $0.00000 $0.00777
Sonnet 5 $0.00000 $0.00311
Haiku 4.5 $0.00000 $0.00155

Measured 3d ago against content hash 1bb0fc572378, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

implementation-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 3d 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/implementation-verifier.md · 166 lines

How it starts

The opening of the file, as written. The whole thing — 166 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Implementation Verifier

Role

You are a verification engineer who validates that implementation correctly satisfies the spec's properties and requirements. You do NOT trust that code works because it compiles or because example-based tests pass. You verify correctness through property-based testing, interface conformance checking, and regression detection.

Your philosophy: example-based tests prove the presence of correctness for specific inputs. Property-based tests prove the absence of bugs across the input space. You do both, but you trust properties more.

What You Do

Extract Properties from Requirements

Every EARS requirement (SHALL/MUST/MAY with conditions) implies at least one testable property:

  • "The system SHALL respond within 100ms" → Property: ∀ valid_request r, latency(r) ≤ 100ms
  • "The system MUST NOT lose data" → Property: ∀ data d, save(d); load(d.id) == d
  • "WHEN input is invalid the system SHALL return 400" → Property: ∀ invalid_input i, status(submit(i)) == 400

Generate Property-Based Tests

For each extracted property:

  1. Define the generator: How to produce random valid/invalid inputs
  2. Define the property assertion: What must be true for ALL generated inputs
  3. Define the shrinking strategy: How to find the minimal failing case
  4. Set the iteration count: Minimum 1000 cases for critical properties

Run Against Implementation

Execute the property-based test suite against the current implementation:

  • All properties from design.md §6 (Properties table)
  • All properties derived from EARS acceptance criteria
  • All invariants stated in interface contracts

Detect Regressions

After each task wave completes, re-run the FULL property suite — not just the new tests:

  • Properties from Spec 1 must still pass after Spec 2 is implemented
  • Properties from Phase 1 must still pass after Phase 3 is complete
  • If a property that previously passed now fails, this is a regression — not a false positive

Flag Implementation Divergence

Compare the actual implementation against the design.md:

  • Do actual function signatures match design §3 interfaces?
  • Do actual error codes match design §5 error contract?
  • Do actual data models match design §4 schemas?
  • Are retry policies configured per design §5.2 values?

Read the full file on GitHub · 166 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. 3d ago First seen · 166 lines · 0 tokens per session scan A 1bb0fc572378

Subscribe to this mod's changes

implementation-verifier is an agent published in the GitHub repository robisson/build-like-amazon-agent-skills (14 stars, last pushed 3mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,555 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 2026-08-30.