Mini-Program Intake: Evidence-First Handoff

Where this fits

You're an agent asked to take over a WeChat mini-program that's been idle for months. The README says Vue 2, but the lockfile shows Vite. The previous dev's handoff notes mention a cloud function that no longer exists. Without a reliable way to establish ground truth, you risk building on stale assumptions. This skill solves that by enforcing a read-only, evidence-first intake before any planning or modification.

Why agents benefit

• Establishes current facts from the repository itself – scans files, Git state, and configs to build a fact map, ignoring outdated docs.
• Respects a strict read-only boundary – no code changes, dependency installs, or builds, so you can safely inspect any project without side effects.
• Distinguishes current state from historical claims – prevents the classic mistake of treating a past plan as present reality.
• Produces a structured handoff artifact – the fact map becomes a shared reference for you and the user, clarifying risks, unknowns, and change boundaries.
• Works across mini-program flavors – handles WeChat, Alipay, and other mini-program frameworks by parsing common config files and directory patterns.

Practical scenarios

Onboarding onto a legacy mini-program

A user hands you a repo they haven't touched in a year. Instead of guessing, you run intake. You discover the build tool changed, a key page was removed, and the Git branch is behind main. The fact map tells the user exactly what's safe to change and what needs verification.

Auditing before a refactor

Before proposing a migration to TypeScript, you need to know the current component structure and any protected behavior. The intake skill reads all page and component files, notes existing tests, and flags areas with no coverage — giving you a solid baseline for a refactor plan.

Resuming a paused project

A user wants to continue a project that was paused mid-development. Intake reveals uncommitted changes and a half-finished feature. You preserve those changes as user assets and clearly mark what's incomplete, so you can pick up without overwriting work.

Add it to your agent workflow

Install the skill with:

npx skills add NocodeMrLi/mini-program-engineering-skill-suite --skill mini-program-project-intake-skill

Then in your agent prompt, request an intake before any changes:

Run intake on this mini-program project first. I need a fact map of the current state, including framework, rules, Git status, and any risks before we discuss modifications.

The skill will produce a structured report you can use to guide subsequent steps.

Compare before adopting

Before committing to this skill, evaluate it against alternatives. Check the GitHub repo for recent commits and open issues — a dormant project may have outdated patterns. Compare with other mini-program audit skills or manual inspection workflows. Most importantly, verify it fits your typical mini-program stack (WeChat, Alipay, etc.) and that its read-only constraints align with your agent's safety requirements. See the full skill details at OpenAgentSkill for a deeper dive.

Why it is worth tracking

This skill addresses a real pain point: agents often inherit messy, undocumented mini-program projects. Its 22 stars are modest but indicate early community validation, and the author's engineering-suite approach suggests ongoing development. Track it if you frequently handle mini-program takeovers or need a reliable, non-destructive way to scope unknown codebases. Re-evaluate when the repo shows active maintenance or when you need a similar intake pattern for other project types.