yeachan-heo/oh-my-claudecode/drydock

Claude Code and agent skill by yeachan-heo/oh-my-claudecode

Trust tier 1
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Lay the keel of the shipyard harness in any repo — the 4-pillar shared environment (Context, Rules, Tools, Standards), seeded as two faces on day one (CLAUDE.md + GLOSSARY.md) with every other surface appearing on first use, so that every human and agent inherits the same design language and anyone can ship. Run once per repo; re-run with --check to audit drift.

Install

npx vskill@latest install yeachan-heo/oh-my-claudecode/drydock

SKILL.md

The same SKILL.md was published earlier as yeachan-heo/oh-my-claudecode/drydock.

Drydock

Lay the keel of the shipyard: one repo, one shared harness, every contributor inherits it. This skill seeds exactly two faces (CLAUDE.md and GLOSSARY.md), with a read-once migration for recognizable legacy glossary CONTEXT.md files, and defines the first-use moment when every other surface appears. It wires the flows that fill them (launch writes GLOSSARY.md vocabulary and decision records; launch C5 sediment and reviews sediment standards) and reports what was seeded, what is deferred and why, and what was skipped on purpose.

The four pillars and where they physically live:

PillarSurfaces
Context (shared background)GLOSSARY.md (glossary) + docs/business/ + docs/adr/ + OMC wiki
Rules (boundaries)CLAUDE.md (thin entry: conventions, principles, index) + docs/standards/
Tools (composable capability).omc/skills/ + .mcp.json + scripts/
Standards (the classification society)design-system/ (tokens, components, patterns) + docs/standards/

Metaphor map: the shipyard is the shared facility; the classification society (docs/standards/ + design-system/) sets the rules a ship must pass to be seaworthy; drydock lays the keel; launch ships it.

When to Use

  • starting a repo that humans and agents will both build on
  • a repo where knowledge lives in people's heads and chat history instead of files
  • onboarding: a new teammate or agent should inherit context by reading, not by asking

When Not to Use

  • throwaway prototypes with no collaborators
  • a repo already running this harness (use --check instead)

Workflow

1. Detect (never clobber)

Inventory what exists before writing anything:

  • CLAUDE.md present? AGENTS.md present? (rule: if either exists, extend it in place; create the missing one as a one-line pointer to the other; never create both fresh)
  • GLOSSARY.md (or a recognizable legacy glossary CONTEXT.md), docs/adr/, docs/standards/, docs/business/, design-system/, .omc/skills/, .mcp.json, scripts/, .gitattributes — which exist (adopted — their integrity is audited), which are absent (not adopted — each gets its first-use trigger below, never a speculative creation)?
  • OMC installed? — only worth checking when running inside an OMC session; outside one, skip this check silently (the harness works with or without OMC)

Report the map first, then act.

2. Resolve document language, then ask only what detection cannot answer

The document language for the generated harness files is a file-backed decision, not conversation state. Use this contract exactly:

<!-- shipyard-document-language-contract:start -->

json
{
"schemaVersion": 1,
"authority": { "path": "GLOSSARY.md", "frontmatterKey": "documentLanguage" },
"canonicalSources": ["CLAUDE.md", "README.md"],
"askOn": ["missing", "mixed", "conflict", "low-confidence", "invalid-explicit", "script-ambiguous"],
"tagPattern": "^[a-z]{2,3}(?:-[A-Z][a-z]{3})?(?:-(?:[A-Z]{2}|[0-9]{3}))?$",
"scriptVariants": ["zh-Hans", "zh-Hant"],
"seedCompanionPrefixes": { "en": "en", "zh-Hans": "zh-Hans", "zh-Hant": "zh-Hant" },
"stableTokens": [
"GLOSSARY.md", "documentLanguage", "/oh-my-claudecode:launch", "--serial",
"plan", "execute", "review", "verify", "blockedBy", "blocked_by",
"pending", "in_progress", "completed", "failed", "ready-for-agent",
"id", "name", "description", "triggers", "mcpServers", "```",
"<Project>", "<term>", "<feature-slug>"
]
}

<!-- shipyard-document-language-contract:end -->

Legacy glossary migration (full runs only; --check never renames or rewrites anything and just audits the legacy file in place): a CONTEXT.md is a legacy glossary only when its YAML frontmatter carries a documentLanguage key or its first visible non-blank line (after the frontmatter, ignoring a BOM and HTML comments) is a drydock glossary seed heading (Glossary, 术语表, 詞彙表 — seeds that predate the language tag); any other CONTEXT.md is an unrelated agent-context document and is left alone. If the yard has a legacy glossary CONTEXT.md but no GLOSSARY.md, rename it to GLOSSARY.md (git mv when tracked) before resolving the language or scaffolding anything, preserving its content and frontmatter byte-for-byte. Never create a fresh GLOSSARY.md beside a legacy glossary CONTEXT.md; one yard has exactly one glossary authority. An unrelated CONTEXT.md does not block creating GLOSSARY.md and is never reported for removal. If both GLOSSARY.md and a legacy glossary CONTEXT.md exist, GLOSSARY.md is authoritative and the leftover CONTEXT.md is reported for the human to merge or remove.

CLAUDE.md rewrite during migration: Whenever GLOSSARY.md is the glossary authority (after the rename above, and also when GLOSSARY.md and a legacy glossary CONTEXT.md both exist), rewrite legacy CLAUDE.md references such as Glossary: CONTEXT.md or On term conflicts CONTEXT.md wins to GLOSSARY.md, so installed instructions never point at a file that is renamed or about to be removed. The rewrite belongs to the migration, not the scaffold step.

Resolution order:

  1. An explicit human choice in the current invocation wins when valid. Normalize it to a stable BCP-47-style tag: lowercase language, Title-Case script, uppercase region. Invalid explicit input must be asked once rather than guessed.
  2. Otherwise, read documentLanguage from the YAML frontmatter at the top of GLOSSARY.md. A valid, script-unambiguous tag is authoritative for fresh Drydock and Launch invocations. If the persisted tag is bare or region-only Chinese, ask once at this authority tier; never bypass it with source inference.
  3. If the marker is absent or invalid, inspect canonical sources in this order: CLAUDE.md, then README.md. Infer only when every usable source has one unambiguous dominant language and all usable sources agree on the same normalized tag. One unambiguous source is sufficient when the other is missing or empty.
  4. Chinese must resolve to an explicit script-qualified tag: zh-Hans or zh-Hant (optionally followed by a region). Bare zh and region-only Chinese tags are script-ambiguous and must be asked once rather than selecting a companion. Companion selection uses the longest language/script prefix: zh-Hans-* selects the zh-Hans companion and zh-Hant-* selects zh-Hant; preserve the full normalized tag (for example zh-Hans-CN) in GLOSSARY.md.
  5. Missing usable sources, mixed-language content, conflicting tags, low-confidence inference, invalid explicit input, or script-ambiguous Chinese must trigger one batched language question. Do not guess. If no answer is available, stop before writing localized artifacts.
  6. Before scaffolding, write the resolved tag to the exact stable frontmatter key documentLanguage in GLOSSARY.md (creating or extending its frontmatter without translating the key). This visible file is the init report's language authority; no daemon, hidden ledger, or runtime state is created.

Only prose and human-facing labels/localizable values follow the selected language; structural keys stay language-stable. Keep paths, slash commands, flags, code fences, placeholders, frontmatter keys and machine-semantic values, YAML/JSON keys, lifecycle tokens, status enums, IDs, blockedBy, public Team blocked_by, and parser/control tokens byte-for-byte stable.

Ask the remaining questions only after language is resolved:

  • package/tech stack (for standards and design-system seeds)
  • does this repo have a UI? (informational only — design-system/ follows the same deferred-surface rule as everything else: it is created when the first UI token or component contract is worth reusing, never as a stub)
  • issue tracker location (GitHub — V1 supports GitHub only, no local fallback) — also record the maintainer's communication authority for intake (issue comments, label changes) — consumed by the navigator's map home and the harbor's intake queue
3. Scaffold (seed the two faces — seeds render in the document language)

Drydock writes exactly two faces on day one:

CLAUDE.md # thin entry — see seed A
GLOSSARY.md # glossary and language authority — see seed B

Every other surface is deferred: it appears on first use, never speculatively. A surface exists to hold something; it is created the first time that something actually happens, seeded with the matching exemplar below. A file created with no first use behind it is session content pretending to be a repository surface:

SurfaceAppears when
docs/adr/the first load-bearing decision settles — write ADR-0001 at that moment (adopting the harness is the natural first entry)
docs/standards/the first checkable rule is sedimented — a launch C5 sediment pass or a review correction that repeats (seeds C / C2)
docs/business/the first business rule or background article must outlive the conversation (seed D)
design-system/the first UI token or component contract is worth reusing — UI repos only (seed E)
.omc/skills/the first reusable capability passes the skillify gate (seed F)
scripts/the first automation is needed more than once
.mcp.jsonthe first tool integration is actually wired (seed: {"mcpServers": {}} — servers get added when a tool integration is actually needed, not speculatively)
.gitattributesthe repo first hits line-ending churn (seed: * text=auto eol=lf — kills CRLF warning noise on Windows)
yard audit script (scripts/shipyard-audit.mjs)the governance loop is first exercised — a launch yard gate or --check drift audit needs the mechanical half; seed it at that moment

Seed exemplars are reference companions, never a combined payload. Select exactly one companion after resolving documentLanguage; do not emit duplicate headings or labels from another companion. Use the longest matching language/script prefix: en-* uses English, zh-Hans-* uses Simplified Chinese, and zh-Hant-* uses Traditional Chinese, while Seed B writes the full resolved tag into documentLanguage. For any other valid tag, translate the English canonical companion once while preserving every stable token above. Seeds A and B are written now; seeds C–F wait for their surface's first-use trigger and render in the document language when that moment comes.

Before generating seed prose, call the Skill tool with agent-doc-discipline and apply its rules; seed prose is ready only when every rule is checkable and carries a why, every surface is self-describing without chat history, and sources are named rather than assumed. A teammate or agent should be able to act on the seed's content by reading alone.

Seed A — CLAUDE.md, en (thin entry; extend in place if the file exists):

<!-- shipyard-seed-a:en:start -->

markdown
# <Project> — Agent & Human Shipyard
## Project conventions
-<language/framework/package manager/naming — list what matters, skip the rest>
## Architecture principles
-<the 3-5 principles most often violated in this project>
## Shared background
-Glossary: GLOSSARY.md
## Agent guide
-Delivery follows the canonical workflow plan → execute → review → verify; `/oh-my-claudecode:launch` is an optional governed delivery pipeline (opt-in, invoke explicitly)
-On term conflicts GLOSSARY.md wins; new terms are recorded the moment they settle
-When a deferred surface first appears (standards, decision records, business knowledge, project skills, the design system), it gets an index section here at the moment it is created — never before

<!-- shipyard-seed-a:en:end -->

Seed A — zh-Hans companion (结构一致,二选一按文档语言渲染):

<!-- shipyard-seed-a:zh-Hans:start -->

markdown
# <Project> — Agent & Human Shipyard
## 项目约定
-<language/framework/package manager/naming — list what matters, skip the rest>

This is the start of the file. Read the full SKILL.md on GitHub.

Security

  • Security scanPassed, 95 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

Repository stars
39,257
Repository forks
3,514
Repository npm downloads, 7 days
4,835
7-day trend score
72

Stars and forks are for the whole GitHub repository, not this skill alone. Downloads are for the repository's npm package. Refreshed 3 Oct 2026, a month ago.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/yeachan-heo/oh-my-claudecode/drydock/badge)](https://verified-skill.com/skills/yeachan-heo/oh-my-claudecode/drydock)

Something wrong with this skill or its page? Tell us.