Claude Code Plugins

Community-maintained marketplace

Feedback

pre-flight-check

@majiayu000/claude-skill-registry
27
0

Checks essential information and project status before starting a task.

Install Skill

Shared

Installs to .agents/skills, used by Codex, Amp, Warp, Cursor, OpenCode, and more.

CodexAmp
WarpCursorOpenCode
ClineGemini CLI
GitHub Copilot
Personal

Available across projects.

$npx skills-installer add @majiayu000/claude-skill-registry/pre-flight-check --client shared
Project

Writes to .agents/skills.

$npx skills-installer add @majiayu000/claude-skill-registry/pre-flight-check -p --client shared
Note: Review the skill instructions before using it.

SKILL.md

name pre-flight-check
description Checks essential information and project status before starting a task.
context fork

Pre-Flight Check Skill

Role: Check essential info and project status before starting work to reduce omissions.

Inputs

  • Feature name/branch name (optional)
  • Required doc paths: CLAUDE.md, context.md, etc.

Checklist

  • UI spec version/design assets availability
  • API spec availability
  • Similar feature references
  • git status/branch, build status
  • context.md freshness, unresolved items in pending-questions.md
  • Document Memory Policy check:
    • context.md token usage (warn if > 6,000 tokens, ~80% of limit)
    • specification.md exists and is summarized (if large spec)
    • archives/ directory structure in place

Output (example)

# Pre-flight Check Results

## Required Info
OK UI spec: v3 (YYYY-MM-DD)
OK API spec: draft available
WARN similar feature reference: not found

## Project Status
OK git status: clean
OK branch: feature/{feature-name}
OK build status: success

## Docs
OK CLAUDE.md: latest
WARN context.md: missing (needs creation)

## Document Memory Policy
OK context.md tokens: ~3,200 (under 8,000 limit)
OK specification.md: summarized (full in archives/)
OK archives/ directory: exists

## Recommended Actions
1. [HIGH] Create context.md (ContextBuilder Agent)
2. [MEDIUM] Verify design assets (invoke design-spec-extractor)

Anti-Pattern Check

Detect common mistakes based on "How to write a good spec for AI agents" guide

Checklist

Anti-Pattern Detection Method Status
Vague prompts "Make something cool", "Make it work better" ⚠️
Large context without summary specification.md > 3,000 tokens + no summary ⚠️
Missing 6 core areas PROJECT.md lacks commands/tests/structure/style/Git/boundaries ⚠️
Vibe coding ↔ Production confusion Attempting production deploy without tests ⚠️
Ignoring deadly trinity Proceeding without verification for speed/non-determinism/cost ⚠️

Recommended Actions on Detection

Detected Anti-Pattern Recommended Action
Vague request Invoke requirements-analyzer to clarify
Large document Follow document-memory-policy.md to summarize
Missing areas Recommend PROJECT.md review, reference template sections
No tests Recommend writing tests before completion-verifier

Output Example

## Anti-Pattern Check Results

| Item | Status | Action |
|------|--------|--------|
| Clear request | ✅ | - |
| Context size | ✅ | ~2,500 tokens |
| 6 core areas | ⚠️ | Git workflow undefined |
| Test defined | ✅ | - |

**Recommended Actions:**
1. [MEDIUM] Add Git Workflow section to PROJECT.md

References

  • .claude/docs/guidelines/document-memory-policy.md