Spark Program | CCC Vibe-Coding Scaffold (AI-Assisted Dev Starter Kit)
Project Name
CCC Vibe Scaffold — A CCC project template + test harness designed for safe AI-assisted (“vibe-coded”) development
This proposal responds directly to a live, named discussion in the Spark Program’s own 2025 Annual Report thread — we’re not guessing at a need, we’re building the thing community members and committee members were actively discussing wanting.
Problem
In the discussion under the Spark Program’s 2025 Annual Report, a community member (janx) raised that AI-assisted (“vibe-coded”) development is now roughly 100x faster for application-layer work, and proposed the program actively support it. The Spark Committee (zz_tovarishch) responded positively but with two explicit guardrails: (1) disclosure of what AI tools were used for which parts, and (2) keeping critical on-chain logic manual-coded or audited while allowing aggressive AI acceleration off-chain. The committee explicitly floated “templates, CCC-based scaffolds, test harnesses, and verifiable-by-default workflows” as a future tooling category — but nobody has built one yet.
Right now, a developer who wants to vibe-code a CKB dApp has no starting scaffold that already encodes these guardrails. They’d have to invent the disclosure format and the on-chain/off-chain safety boundary themselves, every time.
Solution
CCC Vibe Scaffold is a small, open-source project template (a “create-ckb-app”-style starter) that:
Ships with a CCC-based project structure that clearly separates off-chain logic (safe to heavily AI-generate) from on-chain script logic (flagged for manual review/audit), matching the boundary the committee itself described.
Includes a built-in test harness that runs automatically against any off-chain code changes, so AI-generated code has an immediate, objective pass/fail signal rather than relying on manual review. The harness uses [Vitest/Jest — confirm which] and covers unit and integration tests against off-chain logic (transaction building, CCC calls). On-chain script logic is intentionally out of scope for automated testing and stays flagged for manual review, per the committee’s original guardrail. Tests trigger automatically via a pre-commit hook and CI workflow on any off-chain code change — no manual step required.
Includes a simple, standard “AI Disclosure” manifest — a YAML file (.ai-disclosure.yml) at the project root, chosen specifically because it needs to stay human-editable. A solo vibe-coding developer can update it by hand in a couple of lines per file rather than running a generator. Example:
yaml
files:
- path: src/mint.ts
ai_tool: Claude
scope: "full generation, off-chain logic"
- path: contracts/lock.rs
ai_tool: none
scope: "manual, audited"
Ships with a worked example dApp (a minimal transfer/mint flow) built using the scaffold, demonstrating the pattern end-to-end, including its own filled-out disclosure manifest.
Relevance to the CKB Ecosystem
Directly implements a tooling category the Spark Committee named explicitly, in writing, in their own forum thread — this is about as direct a response to stated demand as a Spark application can be.
Supports CCC adoption (a named 2026 priority) by lowering the time-to-first-dApp even further for AI-assisted builders.
Produces a reusable disclosure format that, if adopted, could become a small piece of ecosystem-wide convention — useful well beyond this one grant.
Expected Deliverables
Open-source scaffold/template repository (MIT or Apache-2.0), usable via a single scaffolding command. Built-in automated test harness wired to the off-chain/on-chain split. AI Disclosure manifest format (YAML) + template file. One worked example dApp built on the scaffold, with its own disclosure manifest filled out as a real example. README documenting the on-chain/off-chain boundary convention and how to use the disclosure format.
How to Verify
Run the single scaffolding command from the README to generate a new project — reviewers confirm it produces a working CCC-based project structure with no manual fixes needed. Run the included test harness against the worked example dApp — it should pass cleanly, demonstrating the test-harness-as-safety-net pattern actually works. Open the worked example’s AI Disclosure manifest and cross-check it against the actual commit history in the repo, confirming the disclosure accurately reflects what was AI-assisted versus manually written.
Requested Funding
$1,000 USD, 100% CKB.
Scaffold/template structure and scaffolding CLI: $400 — covers the project generator and the on-chain/off-chain folder convention. Test harness: $300 — covers the automated pass/fail signal for off-chain code. AI Disclosure format and worked example dApp: $300 — covers the manifest template and one filled-out real example.
Estimated Timeline
4 weeks total. Week 1: scaffold structure and scaffolding CLI working. Week 2: test harness wired in and passing on a trivial project. Week 3: AI Disclosure manifest format finalized and worked example dApp built. Week 4: docs, polish, buffer for committee feedback.
To-Do List
Build scaffold/template structure. Build scaffolding CLI. Implement automated test harness. Design AI Disclosure manifest format. Build worked example dApp using the scaffold. Fill out real disclosure manifest for the example. Documentation. Submit completion report.