scaffold

A project setup workflow that interviews the user and creates technical documentation, architecture notes, project instructions, and a phased implementation plan.

In plain words
What is it for?
Use it to define a project’s technology choices, architecture, documentation, coding roles, and implementation phases.
Why use it?
It turns an empty or existing codebase into a documented project with decisions and work organized before implementation starts.

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/ccsk-org/ccsk-cli/scaffold
Clone the repo
git clone --depth 1 https://github.com/ccsk-org/ccsk-cli

Made for: Claude Code.

Per session 44 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 3,143 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. 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 $0.00044 $0.03143
Opus 5 $0.00022 $0.01571
Sonnet 5 $0.00009 $0.00629
Haiku 4.5 $0.00004 $0.00314

Measured 2d ago against content hash 04b2f12083c7, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

scaffold scanned grade A with 1 finding 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.

Runs shell commandslowCapability

Expected in a hook, worth knowing in a rule or an instructions file.

- `dangerouslySetInnerHTML`, `eval(`, `child_process.exec(` with concatenation.
.claude/commands/scaffold.md · 234 lines

How it starts

The opening of the file, as written. The whole thing — 234 lines — stays where its author put it; the contents beside it link to each section on GitHub.

scaffold

Canonical name: scaffold. Lives at .claude/commands/scaffold.md. One command that takes the user from "fresh kit installed" to a documented, planned, ready-to-implement project. This file is the complete master prompt. Follow it literally.


Pre-flight rules (non-negotiable)

  1. Don't assume. Don't guess. Surface concerns. Anything not derivable from the codebase or the user's words — ask.
  2. Use AskUserQuestion for every decision the user owns. Single-text prompts only when free-form input is essential (project name, one-line goal).
  3. If a step is genuinely impossible, name it and stop. Never fabricate. Never cheat to keep the flow moving.
  4. Token discipline. Read files in slices. Use Glob/Grep before Read. Don't dump full files into your reply.
  5. Knowledge cutoff awareness. Your training data is bounded. For any concrete version number, framework feature, or "current latest" claim you write into a file, you MUST verify via context7 (preferred) or docs-seeker skill before writing it down. If you cannot verify, write the category (e.g. "App-Router meta-framework") and mark it TODO: pin version after verification.
  6. CLAUDE.md edits are surgical. You may rewrite ## Role & Responsibilities and ## Project Overview (or create them if missing) only when the content materially differs from what's there. Phrasing-only diffs → leave them alone. Every other heading is read-only for this command.
  7. No new tooling without consent. Don't install packages, init git repos, or run codegen as a side-effect of bootstrap. This command writes markdown only.

Step 0 — Detect mode

Run, in order, then choose:

test -f package.json && echo HAS_NODE
test -f pyproject.toml && echo HAS_PY
test -f Cargo.toml && echo HAS_RUST
test -f go.mod && echo HAS_GO
ls -A | grep -Ev '^(\.git|\.claude|\.ccsk|docs|README\.md|CLAUDE\.md|LICENSE|\.gitignore)$' | head -5
git rev-parse --is-inside-work-tree 2>/dev/null

Read the full file on GitHub · 234 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. 2d ago First seen · 234 lines · 44 tokens per session scan A 04b2f12083c7

Subscribe to this mod's changes

scaffold is a command published in the GitHub repository ccsk-org/ccsk-cli (55 stars, last pushed 1mo ago), licensed MIT. It adds 44 tokens to every session and 3,143 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 1 finding (runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.