openspec-check-standards

Check code against project development standards defined in openspec/context/project_summary.md.

Input: Optionally specify file paths, directories, or a change name. If omitted, checks recently modified files or files in active change.

Steps

1. •

   Load development standards

   Read openspec/context/project_summary.md and extract the "Development Standards" section.

   Key rules to check:

   Data Processing (Frontend):

   • •No frontend pagination (client-side slicing of full data)
   • •No frontend sorting (Array.sort on API data)
   • •No frontend filtering/searching (filter on full dataset)
   • •No pseudo-pagination (loading all data, showing pages)

   API Design:

   • •List APIs must support page + page_size parameters
   • •List APIs must return total_count
   • •No page_size > 100
   • •No SELECT without LIMIT

   Frontend Constraints:

   • •Must show Loading/Empty/Error states
   • •API calls must go through React Query hooks

   Backend Constraints:

   • •List queries must have default limit
   • •Must use parameterized queries
   • •Slow queries (>1s) must be logged
2. •

   Determine scope

   If change name provided:

   • •Get files modified in that change from openspec/changes/<name>/tasks.md
   • •Or check git diff if tasks reference file paths

   If file/directory paths provided:

   • •Use those paths directly

   If nothing provided:

   • •Check recently modified files (git status)
   • •Or scan frontend/src/ and backend/app/ for common issues
3. •

   Run checks

   For each file in scope:

   Frontend Checks (*.tsx, *.ts in frontend/)

   Anti-pattern: Frontend Pagination

   code

   Search for patterns like:
   - .slice(startIndex, endIndex)  // on API response data
   - data.slice(page * pageSize    // pseudo-pagination
   - useState.*page.*setPage       // local pagination state with no API call
   

   Anti-pattern: Frontend Sorting

   code

   Search for patterns like:
   - .sort((a, b) =>               // sorting API response
   - [...data].sort(               // copying and sorting
   - useMemo.*sort                 // memoized sort on full data
   

   Anti-pattern: Frontend Filtering

   code

   Search for patterns like:
   - .filter(item =>               // on API response data
   - data.filter(                  // filtering full dataset
   - useMemo.*filter               // memoized filter on API data
   

   Required: Loading States

   code

   Check for:
   - isLoading / isPending usage
   - Skeleton / Spinner components
   - Loading fallback UI
   

   Required: React Query

   code

   Check for:
   - useQuery / useMutation usage
   - No direct fetch() in components
   

   Backend Checks (*.py in backend/)

   Anti-pattern: Missing LIMIT

   code

   Search for patterns like:
   - .all() without .limit()
   - SELECT * FROM without LIMIT
   - query.all() on unbounded queries
   

   Anti-pattern: Missing Pagination Params

   code

   Check router functions for:
   - List endpoints missing page/page_size params
   - Missing skip/limit in queries
   

   Required: Structured Logging

   code

   Check service classes for:
   - logger.info/warning/error usage
   - Key fields: task_id, duration, retry_count
   

4. •

   Generate report

   code

   ## Standards Check Report
   
   **Scope:** [files checked]
   **Date:** [timestamp]
   
   ### Summary
   | Category | Passed | Failed | Warnings |
   |----------|--------|--------|----------|
   | Frontend Data Processing | X | Y | Z |
   | API Design | X | Y | Z |
   | Backend Constraints | X | Y | Z |
   
   ### Issues Found
   
   #### CRITICAL (Must Fix)
   
   **[File:Line] Frontend Pagination Detected**
   ```tsx
   const paginatedData = data.slice(page * 10, (page + 1) * 10);
   

   • •Problem: Loading full data then slicing client-side
   • •Fix: Add page and page_size params to API, use server pagination

   WARNING (Should Fix)

   [File:Line] Missing Loading State

   • •Problem: No loading indicator while fetching
   • •Fix: Add {isLoading && <Spinner />} or skeleton

   SUGGESTION (Nice to Fix)

   [File:Line] Consider Structured Logging

   • •Add task_id and duration to log entries

   Passed Checks

   • •✓ React Query usage
   • •✓ Parameterized queries
   • •✓ Default LIMIT on queries
   code

5. •

   Provide actionable next steps

   Based on findings:

   • •If CRITICAL issues: "Fix these before proceeding"
   • •If only warnings: "Ready to continue, consider fixing warnings"
   • •If all passed: "All standards checks passed!"

Output

A structured report showing:

• •What was checked
• •Issues found (grouped by severity)
• •Specific file:line references
• •Concrete fix suggestions
• •Overall status (PASS/WARN/FAIL)

Integration with Other Skills

• •Run after openspec-apply-change completes tasks
• •Run before openspec-verify-change for complete validation
• •Run before openspec-archive-change as final gate

Guardrails

• •Don't report false positives for legitimate use cases (e.g., local UI state sorting)
• •Distinguish between API data manipulation and UI-only data
• •If uncertain, mark as SUGGESTION not CRITICAL
• •Always provide file:line references for issues
• •Focus on patterns, not exhaustive analysis
• •Keep the report actionable and concise