verify

A command that reviews whether completed code changes follow a feature specification. It finds a suitable recent specification, confirms which one to inspect, and sends the review to a specialized verification agent.

In plain words
What is it for?
Use it after a specification has been implemented when you want to check compliance. You can name the specification directly, or let the command look for recently implemented specifications.
Why use it?
It gives a defined final check after implementation instead of relying only on the person who wrote the code. The review focuses on requirements, scope, affected modules, and the specified tests.

Command for Claude Code

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 commands/heiko-braun/draft/verify
Clone the repo
git clone --depth 1 https://github.com/heiko-braun/draft

Made for: Claude Code.

Per session 19 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 483 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 $0.00019 $0.00483
Opus 5 $0.00010 $0.00242
Sonnet 5 $0.00004 $0.00097
Haiku 4.5 $0.00002 $0.00048

Measured 2d ago against content hash 44f5ccbb1473, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

verify 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.

.claude/commands/verify.md · 58 lines

What it actually says

Spec Verification Command

You coordinate the verification of implemented code changes against their specification by delegating the work to a specialized subagent.

When to Activate

  • When user says /verify or "verify the implementation"
  • After implementation is complete and user wants to verify spec compliance
  • When user wants to check if code matches a specific spec

Workflow

1. Identify Spec to Review

Check for user input:

  • If the user provided arguments (e.g., /verify feature-name), use that spec file
  • If no arguments, look for recently implemented specs:
    • List all files in specs/ directory
    • Check their status field in the front-matter
    • Look for specs with status: implemented or recent modifications

Ask user to confirm:

  • Present the identified spec and ask: "I will review the implementation of '{spec-name}'. Is this correct?"
  • If user says no, ask: "Which spec would you like me to review?"

2. Invoke Verification Subagent

Launch the subagent:

When executing this skill, invoke a subagent via the Anthropic API with the following setup:

  • System prompt: .claude/agents/verify-agent.md
  • Pass the spec file path as the user message
  • Parse the subagent's response and use it to complete the task

3. Handle Fixable Issues

After receiving the verification report:

  • Display the full report to the user
  • If the report includes a "Fixable Issues" section with one or more issues:
    • Ask the user: "Would you like me to fix these issues?"
    • If user says yes or confirms:
      • Work through each fixable issue one by one
      • Use appropriate tools (Edit, Write, Bash) to implement the fixes
      • After all fixes are applied, re-run the test suite
      • Report the results to the user
    • If user says no or declines:
      • End the verification process
  • If no fixable issues are found:
    • End the verification process after displaying the report
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 · 58 lines · 19 tokens per session scan A 44f5ccbb1473

Subscribe to this mod's changes

verify is a command published in the GitHub repository heiko-braun/draft (5 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 19 tokens to every session and 483 once invoked, about $0.0001 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.