spec-improve

spec-improve is a command for coding agents from mrvladd-d/memobank. It costs 16 tokens per session (1,344 once invoked), scanned A, original, MIT.

A command that improves the technical design for one feature by checking it against the project's design specifications.

In plain words
What is it for?
It reads the feature, linked requirements, shared design documents, and design index, then updates only the necessary design records and reports the feature's design status.
Why use it?
It finds missing details and contradictions before a feature is divided into implementation tasks.

Command

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/mrvladd-d/memobank/spec-improve
Clone the repo
git clone --depth 1 https://github.com/mrvladd-d/memobank

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 spec-improve

README.md
[![agentmods](https://agentmods.dev/badge/commands/mrvladd-d/memobank/spec-improve.svg)](https://agentmods.dev/commands/mrvladd-d/memobank/spec-improve)
Your own site
<a href="https://agentmods.dev/commands/mrvladd-d/memobank/spec-improve"><img src="https://agentmods.dev/badge/commands/mrvladd-d/memobank/spec-improve.svg" alt="Measured on agentmods" height="20"></a>
Per session 16 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,344 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 $0.00016 $0.01344
Opus 5 $0.00008 $0.00672
Sonnet 5 $0.00003 $0.00269
Haiku 4.5 $0.00002 $0.00134

Measured 5d ago against content hash ca634f128677, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

spec-improve 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 5d 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.

skills/_shared/references/commands/spec-improve.md · 133 lines

How it starts

The opening of the file, as written. The whole thing — 133 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/spec-improve - Feature-level SDD design improvement

/spec-improve FT-<NNN> checks existing specs first, finds gaps/contradictions, asks focused questions when needed, updates only necessary design artifacts, and marks the target feature with spec_design_status.

0) Input

Expected $ARGUMENTS:

  • FT-<NNN>

Run after /prd and mandatory /spec-design, before /prd-to-tasks FT-<NNN>. Consume backbone specs as normative inputs instead of duplicating them.

If the argument is missing, ask the user to choose one feature.

Examples:

  • /spec-improve FT-001
  • /spec-improve FT-012

Canonical route: /write-prd -> /spec-init -> /prd -> /spec-design -> /spec-improve FT-<NNN> -> /prd-to-tasks FT-<NNN>.

1) Read existing design surface first

Before creating any new spec:

  1. Read .memory-bank/spec-index.md.
  2. Confirm global backbone status is complete, or minimal with explicit not_applicable areas; if missing, bare minimal, or blocked, stop and route to /spec-design.
  3. Read the target .memory-bank/features/FT-<NNN>-*.md.
  4. Read linked epic, requirements, Constitution, backbone specs, and any existing specs routed by the index.
  5. Search existing .memory-bank/architecture/, .memory-bank/tech-specs/, .memory-bank/contracts/, .memory-bank/domains/, .memory-bank/states/, .memory-bank/adrs/, .memory-bank/testing/, .memory-bank/guides/, and .memory-bank/runbooks/ for overlapping decisions.

Rule: do not create a new spec before checking existing specs through the index. If several features need the same missing domain/contract/state/API/security/data/runtime/testing decision, stop and route to /spec-design instead of creating duplicate feature-local specs. If a task/feature interpretation conflicts with a backbone spec, stop with a blocker instead of choosing locally.

2) Decide required design depth

Classify what the feature needs:

  • none: simple T0/T1-like work with no runtime, contract, state, data, security, migration, or cross-module design impact
  • feature hub only: a small .memory-bank/tech-specs/FT-<NNN>-<slug>.md is enough
  • linked specs: update or create specific architecture/contracts/domains/states/ADR/testing/runbook docs

Read the full file on GitHub · 133 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. 5d ago First seen · 133 lines · 16 tokens per session scan A ca634f128677

Subscribe to this mod's changes

spec-improve is a command published in the GitHub repository mrvladd-d/memobank (55 stars, last pushed 3mo ago), licensed MIT. It adds 16 tokens to every session and 1,344 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 2026-08-30.