mdd

A development workflow called Manual-First Development: Document, Audit, Fix, Verify. It starts new features with documentation and starts fixes with an audit.

In plain words
What is it for?
Use it to build features, audit existing code, check project status, read relevant project and architecture documents, and verify changes after implementation.
Why use it?
It gives feature work and bug fixes a defined sequence, helping the agent understand the project before changing it and check that the result works afterward.

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/dr-code/tessera/mdd
Clone the repo
git clone --depth 1 https://github.com/dr-code/tessera

Made for: Claude Code.

Per session 24 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 3,259 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 $0.00024 $0.03259
Opus 5 $0.00012 $0.01630
Sonnet 5 $0.00005 $0.00652
Haiku 4.5 $0.00002 $0.00326

Measured yesterday against content hash 9d42166de97f, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

mdd 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 yesterday.

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/mdd.md · 391 lines

How it starts

The opening of the file, as written. The whole thing — 391 lines — stays where its author put it; the contents beside it link to each section on GitHub.

MDD — Manual-First Development Workflow

$ARGUMENTS

MDD is the core development workflow. Every feature starts with documentation, every fix starts with an audit. No exceptions.

Step 0 — Detect Mode

Parse $ARGUMENTS to determine the mode:

  • If arguments start with auditAudit Mode (jump to Phase A)
  • If arguments start with statusStatus Mode (jump to Phase S)
  • If arguments are empty → ask the user what they want to do
  • Otherwise → Build Mode (the default — jump to Phase 1)

BUILD MODE — New Feature Development

Phase 1 — Understand the Feature

Read the user's description: $ARGUMENTS

Before writing anything, gather context:

  1. Read CLAUDE.md — understand project rules
  2. Read docs/PROJECT_CONTEXT.md if present — use as primary feature map and quick reference
  3. Read docs/ARCHITECTURE_SUMMARY.md if present — architecture brief
  4. Read only the specific feature doc(s) relevant to this request — do not scan all docs by default
  5. Read docs/ARCHITECTURE.md only if the task changes or crosses architectural boundaries
  6. Read subtree CLAUDE.md files for the areas being touched (e.g., server/CLAUDE.md, client/CLAUDE.md) if present
  7. Scan src/ — only the relevant area for the requested feature

Then ask the user targeted questions using AskUserQuestion. Ask ALL relevant questions upfront in a single interaction — don't spread them across multiple turns:

Always ask:

  • "Does this feature need database storage? If so, what data does it store?"
  • "Does this feature have API endpoints? What operations (create, read, update, delete)?"
  • "Does this feature depend on any existing features?" (list the relevant ones from docs/PROJECT_CONTEXT.md or docs/*.md)
  • "Are there any edge cases or error scenarios you already know about?"

Ask if relevant:

  • "Does this need authentication/authorization?"
  • "Does this need real-time updates (WebSocket)?"
  • "Does this need background processing (queues, cron)?"
  • "Does this integrate with any external services?"

Read the full file on GitHub · 391 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. yesterday First seen · 391 lines · 24 tokens per session scan A 9d42166de97f

Subscribe to this mod's changes

mdd is a command published in the GitHub repository dr-code/tessera (1 stars, last pushed 21d ago), licensed MIT. It adds 24 tokens to every session and 3,259 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.