Agent skill

Codd Generate

by yohey-w in yohey-w/codd-dev

Generate CoDD design documents for a greenfield project one wave at a time.

MITAuto-check passedDevelopment

Install Codd Generate

skills CLI
$ npx skills add yohey-w/codd-dev --skill codd-generate -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install yohey-w/codd-dev codd-generate --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/yohey-w/codd-dev.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/codd-generate .claude/skills/codd-generate && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
codd-generate
GitHub stars
115
Token cost
~1.5k tokens
SKILL.md length
734 words
Files
1
Skills in repo
11
Repo updated
First seen
Licence
MIT

At a glance

Generate CoDD design documents for a greenfield project one wave at a time.

  • Works in 3 steps: codd extract — reverse-engineer code… → codd plan --init — generate wave_config… → /codd-restore — reconstruct design docs…
  • Needs requirement-driven design docs
  • SKILL.md covers Usage, Wave Model, Auto Wave Config and Prerequisite Checks, plus 6 more sections
  • Calls python

What it does

Codd Generate is an agent skill from yohey-w/codd-dev. Generate CoDD design documents for a greenfield project one wave at a time. Use after codd init and prepared requirements when the user needs requirement-driven design docs, frontmatter validation, graph refresh, and human approval gates between waves.

Its SKILL.md is about 1.5k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Development, covering Architecture decision records. The repository describes itself as: CoDD: Coherence-Driven Development — Claude Code plugin for cross-artifact change impact analysis. The licence is MIT.

When your agent uses it

  • Needs requirement-driven design docs
  • Frontmatter validation
  • Human approval gates between waves

Example prompts

  • “/codd-generate”

Requirements

  • Python 3

Workflow steps

3 steps, taken from the first numbered list in SKILL.md.

  1. codd extract — reverse-engineer code structure
  2. codd plan --init — generate wave_config from extracted docs
  3. /codd-restore — reconstruct design docs from extracted facts

What it can do on your machine

Read from SKILL.md and the folder at commit 41b5f87. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • python

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Codd Generate loads about 1.5k tokens when it runs. Until then it costs about 67 tokens; SKILL.md has 734 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~67
When it runs · the whole SKILL.md, loaded when a task matches
~1.5k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from yohey-w/codd-dev at commit 41b5f87, republished under its MIT licence (© yohey-w). 734 words, ~1,485 tokens.

Download SKILL.mdSave it as .claude/skills/codd-generate/SKILL.md (or your agent's skills folder).
name
codd-generate
description
Generate CoDD design documents for a greenfield project one wave at a time. Use after `codd init` and prepared requirements when the user needs requirement-driven design docs, frontmatter validation, graph refresh, and human approval gates between waves.

CoDD Generate

Generate CoDD design documents one wave at a time, validate their frontmatter, refresh the dependency graph, and stop for human approval before advancing to the next wave.

This is for greenfield projects (requirements → design). For brownfield projects (existing code → design), use /codd-restore instead.

Usage

Use this skill after codd init and after requirement documents are ready. Generate only one wave at a time. Never auto-run the next wave without a human decision.

If you have an existing codebase with no requirements, use the brownfield flow instead:

  1. codd extract — reverse-engineer code structure
  2. codd plan --init — generate wave_config from extracted docs
  3. /codd-restore — reconstruct design docs from extracted facts

Wave Model

CoDD design generation follows dependency order. Documents in the same wave may be generated together, but later waves must wait until the previous wave is validated and reviewed.

  • Wave 1: requirement-only artifacts such as acceptance criteria and decisions
  • Wave 2: system-level design derived from requirements plus Wave 1 outputs
  • Wave 3: detailed design artifacts derived from approved system design
  • Wave 4+: later artifacts only after the previous wave is validated and approved

Auto Wave Config

Since v0.2.0a4, codd generate automatically generates wave_config from requirement documents if it's missing from codd.yaml. You no longer need to run codd plan --init manually before generating. The flow is:

  1. Write requirements in any format (plain text, markdown, etc.) and import with codd init --requirements <file> — CoDD adds frontmatter automatically
  2. Run codd generate --wave 2 — wave_config is auto-generated from requirements
  3. Design docs are generated in the correct dependency order

This means the only human input is the requirements. Everything else — wave_config, frontmatter, dependency declarations — is derived automatically.

Prerequisite Checks

Before every generation run, verify these conditions:

  1. codd/codd.yaml exists in the project root.
  2. At least one requirement document exists under configured doc_dirs. If imported with codd init --requirements, frontmatter is already added. Otherwise, ensure node_id and type: requirement are present in frontmatter.
  3. For Wave 1, codd init has been completed successfully.
  4. For Wave 2 and later, every output from the previous wave exists and codd validate --path . passes.
  5. Confirm you are in the intended project root before running generation commands.

If any prerequisite fails, stop and resolve it before generating.

Follow this loop for each wave:

  1. Generate the target wave:

    bash
    codd generate --wave 1 --path .

    Replace 1 with the current wave number.

  2. Read every generated document and confirm:

    • the content matches the requirement scope
    • no unsupported features were invented
    • the document body starts directly with content, not AI meta commentary
  3. Validate frontmatter and dependency references:

    bash
    codd validate --path .
  4. Refresh the dependency graph:

    bash
    codd scan --path .
  5. Pause for HITL confirmation before the next wave:

    • After Wave 1: Wave 1の設計書を確認しました。Wave 2に進みますか?
    • After any later wave: Wave Nの成果物を確認しました。Wave N+1 に進みますか?

This skill covers the full generate -> scan -> validate checkpoint set, while the safer operator order is generate -> read -> validate -> scan -> HITL.

Show full SKILL.md (255 more words)Show less

HITL Gates

Human review is mandatory between waves. Do not proceed automatically.

  • Review goal 1: confirm the generated artifacts stay within requirement scope
  • Review goal 2: catch wrong architectural direction before it propagates downstream
  • Review goal 3: approve the next wave explicitly

If the human rejects the wave, revise the current wave first, re-run validation, re-run scan, and request approval again.

Frontmatter Validation Checklist

Run codd validate --path . after each generation and verify:

  • every generated document has CoDD frontmatter
  • node_id and type are present
  • depends_on references resolve to existing upstream nodes
  • wave ordering is consistent with dependency declarations
  • validation exits successfully before you continue

If validation fails, fix the generated document or the wave configuration before moving forward.

Suggested Commands by Wave

Wave 1
bash
codd generate --wave 1 --path .
codd validate --path .
codd scan --path .

Preconditions:

  • codd init has been run
  • codd/codd.yaml exists
  • requirement documents are present
Wave 2+
bash
codd generate --wave 2 --path .
codd validate --path .
codd scan --path .

Preconditions:

  • previous wave outputs exist
  • previous wave passed codd validate --path .
  • human approved progression to the next wave

Troubleshooting

  • Error: codd/ not found. Run 'codd init' first.
    • Run codd init, create codd/codd.yaml, and retry.
  • Validation errors after generation
    • Read the reported file, correct frontmatter or dependency references, then run codd validate --path . again.
  • Generated content drifts beyond requirements
    • discard the unsupported direction, regenerate the same wave with tighter context, and repeat review before advancing.
  • Dependency graph looks stale
    • re-run codd scan --path . after validation completes.

Guardrails

  • Use the codd command, not python -m codd.cli.
  • Generate one wave at a time.
  • Do not skip human approval gates.
  • Do not advance to the next wave while validation is failing.

© yohey-w, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in skills/codd-generate of yohey-w/codd-dev.

Open the folder on GitHubat commit 41b5f87

Compare with similar skills

Codd Generate next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Codd Generate compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Codd Generate this skillyohey-w/codd-dev115—~1.5kAutomated safety check: PassMIT
PR Design DocOpenHands/OpenHands90k—~2.4kAutomated safety check: PassMIT
Cto AdvisorIbrahim-3d/orchestrator-supaconductor3814 repos~2.4kAutomated safety check: PassMIT
Architecture DecisionDonchitos/Claude-Code-Game-Studios26k—~1.7kAutomated safety check: PassMIT
Improve Codebase Architectureywwynm/EverythingDone14415 repos~1.3kAutomated safety check: PassGPL-3.0
Domain Modelingbrim-borium/spotify_sdk1665 repos~806Automated safety check: PassApache-2.0

Similar skills

  • PR Design Doc

    OpenHands/OpenHands

    For a non-trivial pull request, write a self-contained HTML design doc under the temporary .pr/ directory and link a visibility-appropriate preview in the PR description, so maintainers grasp the…

    90k GitHub stars~2.4k tokensUpdated today
    DevelopmentAuto-check passed
  • Cto Advisor

    Ibrahim-3d/orchestrator-supaconductor

    Technical leadership guidance for engineering teams, architecture decisions, and technology strategy.

    381 GitHub starsUsed in 4 repos~2.4k tokens
    DevelopmentAuto-check passed
  • Architecture Decision

    Donchitos/Claude-Code-Game-Studios

    Create an ADR documenting a technical decision: context, alternatives considered, consequences.

    26k GitHub stars~1.7k tokensUpdated yesterday
    DevelopmentAuto-check passed
  • Improve Codebase Architecture

    ywwynm/EverythingDone

    Find deepening opportunities in a codebase, informed by the domain language in CONTEXT.md and the decisions in docs/adr/.

    144 GitHub starsUsed in 15 repos~1.3k tokens
    DevelopmentAuto-check passed
  • Domain Modeling

    brim-borium/spotify_sdk

    Build and sharpen a project's domain model. An agent skill from brim-borium/spotify_sdk.

    166 GitHub starsUsed in 5 repos~806 tokens
    DevelopmentAuto-check passed
  • Design Doc Mermaid

    SpillwaveSolutions/design-doc-mermaid

    Create Mermaid diagrams (flowchart, sequence, class, ER, state, C4, architecture) from text or source code.

    176 GitHub starsUsed in 1 repo~5.6k tokens
    DevelopmentAuto-check passed

More from yohey-w/codd-dev

All 11 skills in this repo
  • Codd Greenfield

    yohey-w/codd-dev

    Run the CoDD greenfield autopilot: a requirements document in, a working system out, unattended.

    115 GitHub stars~2.2k tokensUpdated 24 days ago
    Auto-check passed
  • Codd Init

    yohey-w/codd-dev

    Initialize CoDD in a project and establish the first scan and validation baseline.

    115 GitHub stars~1.1k tokensUpdated 24 days ago
    Auto-check passed
  • Codd Propagate

    yohey-w/codd-dev

    Reverse-propagate source code changes back to affected CoDD design documents.

    115 GitHub stars~1.4k tokensUpdated 24 days ago
    Auto-check passed
  • Codd Restore

    yohey-w/codd-dev

    Reconstruct CoDD design documents from extracted code facts for a brownfield project.

    115 GitHub stars~1.3k tokensUpdated 24 days ago
    Auto-check passed
  • Codd Impact

    yohey-w/codd-dev

    Analyze the downstream impact of changed requirements, design docs, code, or tests in a CoDD project.

    115 GitHub stars~1.5k tokensUpdated 24 days ago
    Auto-check: warnings
  • Codd Scan

    yohey-w/codd-dev

    Refresh a CoDD project's dependency graph from document frontmatter and source code.

    115 GitHub stars~1.4k tokensUpdated 24 days ago
    Auto-check passed

Categories

Questions about Codd Generate

What does Codd Generate do?

Generate CoDD design documents for a greenfield project one wave at a time. Codd Generate is an agent skill from yohey-w/codd-dev. Generate CoDD design documents for a greenfield project one wave at a time.

When should I use Codd Generate?

Codd Generate fits situations like: needs requirement-driven design docs; frontmatter validation; human approval gates between waves.

How do I install Codd Generate in Claude Code?

Run `npx skills add yohey-w/codd-dev --skill codd-generate -a claude-code`. Or copy the skill folder (skills/codd-generate in yohey-w/codd-dev) into .claude/skills/codd-generate in your project. Claude Code loads it when a task matches its description.

How do I install Codd Generate in Codex?

Run `npx skills add yohey-w/codd-dev --skill codd-generate -a codex`. Or copy the skill folder (skills/codd-generate in yohey-w/codd-dev) into .agents/skills/codd-generate in your project. Codex loads it when a task matches its description.

Can I use Codd Generate in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add yohey-w/codd-dev --skill codd-generate -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/codd-generate, .gemini/skills/codd-generate, .github/skills/codd-generate and .opencode/skills/codd-generate in your project.

What does Codd Generate need to run?

Going by SKILL.md and its folder, Codd Generate needs the command-line tools its instructions call (python). Our summary lists: Python 3.

Does Codd Generate access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Codd Generate safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Codd Generate use?

Codd Generate is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Codd Generate use?

About 1.5k tokens (SKILL.md is roughly 5.9k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Codd Generate?

Skills that share tags, products or a category with Codd Generate: PR Design Doc (OpenHands/OpenHands, 90k stars), Cto Advisor (Ibrahim-3d/orchestrator-supaconductor, 381 stars), Architecture Decision (Donchitos/Claude-Code-Game-Studios, 26k stars) and Improve Codebase Architecture (ywwynm/EverythingDone, 144 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Codd Generate?

yohey-w (a GitHub user) maintains it in yohey-w/codd-dev, which has 115 GitHub stars. The repository holds 11 skills in this directory. The repository was last updated on September 14, 2026.

Source: yohey-w/codd-dev on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.