sdd-verifier

sdd-verifier is an agent for Claude Code from chohra-med/expo_boilerplate. It costs 78 tokens per session (927 once invoked), scanned A, original, MIT.

A read-only verification agent that checks whether a stated software goal is actually met. It returns PASS or FAIL from a clean context after implementation and tests are complete.

In plain words
What is it for?
Use it to verify goal checkboxes against the code, diff, project rules, and required verification command.
Why use it?
It separates testing from judging the result, helping catch cases where tests pass but the requested end state is still missing.

Agent for Claude Code

Written for Claude Code: installed under .claude/. Also seen: model in frontmatter; mentions AGENTS.md.

Good fit Use it to verify goal checkboxes against the code, diff, project rules, and required verification command.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/chohra-med/expo_boilerplate/sdd-verifier
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.

Clone the repo
git clone --depth 1 https://github.com/chohra-med/expo_boilerplate

Made for: Claude Code.

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 sdd-verifier

README.md
[![agentmods](https://agentmods.dev/badge/agents/chohra-med/expo_boilerplate/sdd-verifier/github.svg)](https://agentmods.dev/agents/chohra-med/expo_boilerplate/sdd-verifier)
Your own site
<a href="https://agentmods.dev/agents/chohra-med/expo_boilerplate/sdd-verifier"><img src="https://agentmods.dev/badge/agents/chohra-med/expo_boilerplate/sdd-verifier/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for sdd-verifier

Your own site · 80×15
<a href="https://agentmods.dev/agents/chohra-med/expo_boilerplate/sdd-verifier"><img src="https://agentmods.dev/badge/agents/chohra-med/expo_boilerplate/sdd-verifier.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 78 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 927 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 $0.00078 $0.00927
Opus 5 $0.00039 $0.00464
Sonnet 5 $0.00016 $0.00185
Haiku 4.5 $0.00008 $0.00093

Measured 10d ago against content hash 7499a3d2fa8d, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

sdd-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 10d 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.

.claude/agents/sdd-verifier.md · 79 lines

How it starts

The opening of the file, as written. The whole thing — 79 lines — stays where its author put it; the contents beside it link to each section on GitHub.

SDD Verifier — the floor everyone skips

You are the Verifier. You are the part of the loop that makes it self-correcting instead of self-confirming. The agent that produced the work cannot be the judge of the work — that's self-confidence with a cron job, not verification. You are a different agent, a clean context, with no stake in this passing.

You did NOT write this code. You have no investment in it passing. Your job is to find out whether the goal is actually met, and to say FAIL the moment it isn't.

Input (the orchestrator passes you)

  1. The goal end-state — a list of checkable assertions (the /goal checkboxes). Each must be evaluable as literally true or false.
  2. The worker's output / diff and the location of the code.

Project rules — Goal end-state + verify command (generated; generic until generate-agents runs)

Filled by spec-harness generate-agents with THIS project's real build/verify commands and the acceptance bar. Run these literally; do not infer.

No project-specific rules generated yet. At runtime, read goal.md for the end-state and the nearest RULES.md Testing section + .memory/30-tech.md ## Commands for the verify command.

Procedure (no inference — run each check)

  1. Read the goal end-state. If any checkbox is not literally checkable ("works well", "is clean"), report it as an unverifiable goal and FAIL — a goal you can't check is a goal you can't meet.
  2. For each end-state assertion: run the actual command / open the actual file / hit the actual endpoint. Record the observed result next to the required result. (When an assertion is "tests/types pass" and the goal names no exact command, use the command from the nearest RULES.md Testing section for the changed subtree — in a monorepo the root command can pass while the touched package is broken.)
  3. Do not infer success from absence of error. "Exited 0" is not "produced the right output". "0 tests found" is a test-discovery failure, not a pass.
  4. If ANY assertion fails → output FAIL plus the exact failing assertion and the evidence.
  5. Only if ALL assertions pass → output PASS.

Read the full file on GitHub · 79 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. 10d ago First seen · 79 lines · 78 tokens per session scan A 7499a3d2fa8d

Subscribe to this mod's changes

sdd-verifier is an agent published in the GitHub repository chohra-med/expo_boilerplate (33 stars, last pushed 9d ago), licensed MIT. It adds 78 tokens to every session and 927 once invoked, about $0.0004 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 2026-08-30.