Skip to main content
Backend Developmentjeremylongshore

canva-policy-guardrails

'Implement Canva Connect API lint rules, policy enforcement, and automated

Stars
2,267
Source
jeremylongshore/claude-code-plugins-plus-skills
Updated
2026-05-31
Slug
jeremylongshore--claude-code-plugins-plus-skills--canva-policy-guardrails
View on GitHubRaw SKILL.md

// install — copy + paste into any project

mkdir -p .claude/skills && curl -fsSL https://raw.githubusercontent.com/jeremylongshore/claude-code-plugins-plus-skills/HEAD/plugins/saas-packs/canva-pack/skills/canva-policy-guardrails/SKILL.md -o .claude/skills/canva-policy-guardrails.md

Drops the SKILL.md into .claude/skills/canva-policy-guardrails.md. Works with Claude Code, Cursor, and any agent that loads SKILL.md files from .claude/skills/.

Canva Policy & Guardrails

Overview

Automated policy enforcement for Canva Connect API integrations — prevent token leaks, enforce rate limit handling, require error handling, and validate OAuth configuration.

Prerequisites

  • Named owners for OAuth scopes, tenant access, design/asset rights, publication destinations, and retention policy.
  • CI and runtime configuration capable of evaluating policy without exposing live tokens or design content.

Instructions

  1. Enforce reviewed allowlists for scopes, tenant/action combinations, and destinations before issuing an API request.
  2. Fail closed when authorization, provenance, data classification, or required policy configuration is unknown.
  3. Emit only redacted policy decisions and route exceptions through the named approval process.

ESLint Rules

No Hardcoded Credentials

// eslint-rules/no-canva-credentials.js
module.exports = {
  meta: {
    type: 'problem',
    docs: { description: 'Disallow hardcoded Canva OAuth credentials' },
  },
  create(context) {
    return {
      Literal(node) {
        if (typeof node.value !== 'string') return;
        const val = node.value;

        // Canva client IDs start with "OCA"
        if (/^OCA[A-Za-z0-9]{10,}/.test(val)) {
          context.report({ node, message: 'Hardcoded Canva client ID detected. Use environment variable.' });
        }

        // Canva access tokens start with "cnvat_"
        if (/^cnvat_[A-Za-z0-9]{20,}/.test(val)) {
          context.report({ node, message: 'Hardcoded Canva access token detected. Use environment variable.' });
        }
      },
    };
  },
};

Require Rate Limit Handling

// eslint-rules/require-canva-retry.js
module.exports = {
  meta: {
    type: 'suggestion',
    docs: { description: 'Canva API calls should handle 429 responses' },
  },
  create(context) {
    return {
      CallExpression(node) {
        // Check for fetch calls to api.canva.com
        if (node.callee.name === 'fetch' &&
            node.arguments[0]?.value?.includes('api.canva.com')) {
          // Check if parent is try-catch or has .catch()
          const parent = node.parent;
          if (parent.type !== 'AwaitExpression' ||
              parent.parent?.type !== 'TryStatement') {
            context.report({
              node,
              message: 'Canva API calls should be wrapped in try-catch with 429 handling',
            });
          }
        }
      },
    };
  },
};

Pre-Commit Hooks

# .pre-commit-config.yaml
repos:
  - repo: local
    hooks:
      - id: canva-no-tokens
        name: Check for Canva tokens
        entry: bash -c 'git diff --cached --name-only | xargs grep -lE "(cnvat_|OCA[A-Z0-9]{10})" 2>/dev/null && echo "ERROR: Canva credentials found" && exit 1 || exit 0'
        language: system
        pass_filenames: false

      - id: canva-no-raw-urls
        name: Check for hardcoded Canva API URLs
        entry: bash -c 'git diff --cached --name-only | xargs grep -lE "api\.canva\.com/rest/v1" --include="*.ts" --include="*.js" 2>/dev/null | while read f; do grep -n "api\.canva\.com" "$f" | grep -v "const.*BASE\|const.*URL\|import\|from\|//" && echo "WARNING: Direct Canva URL in $f — use client wrapper" && exit 1; done; exit 0'
        language: system
        pass_filenames: false

CI Policy Checks

# .github/workflows/canva-policy.yml
name: Canva Policy Check

on: [push, pull_request]

jobs:
  policy:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4

      - name: Check for hardcoded credentials
        run: |
          if grep -rE "(cnvat_[a-zA-Z0-9]{20,}|OCA[A-Z0-9]{10,})" \
            --include="*.ts" --include="*.js" --include="*.json" \
            --exclude-dir=node_modules .; then
            echo "ERROR: Hardcoded Canva credentials found"
            exit 1
          fi

      - name: Check for unhandled API calls
        run: |
          # Warn if raw fetch to Canva without error handling
          DIRECT_CALLS=$(grep -rn "fetch.*api\.canva\.com" \
            --include="*.ts" --include="*.js" \
            --exclude="*client.ts" --exclude="*test*" \
            . 2>/dev/null | wc -l)
          if [ "$DIRECT_CALLS" -gt 0 ]; then
            echo "WARNING: $DIRECT_CALLS direct Canva API calls found outside client wrapper"
            grep -rn "fetch.*api\.canva\.com" --include="*.ts" --include="*.js" \
              --exclude="*client.ts" --exclude="*test*" .
          fi

      - name: Validate .env.example
        run: |
          if [ -f .env.example ]; then
            for var in CANVA_CLIENT_ID CANVA_CLIENT_SECRET CANVA_REDIRECT_URI; do
              if ! grep -q "^${var}=" .env.example; then
                echo "WARNING: ${var} missing from .env.example"
              fi
            done
          fi

Runtime Guardrails

// Prevent dangerous operations and enforce patterns at runtime

class CanvaGuardrails {
  // Block requests without proper authorization header
  static validateRequest(init: RequestInit): void {
    const authHeader = (init.headers as Record<string, string>)?.['Authorization'];
    if (!authHeader || !authHeader.startsWith('Bearer ')) {
      throw new Error('Canva API call missing Bearer token');
    }
  }

  // Enforce token is not expired before making call
  static validateToken(expiresAt: number): void {
    if (Date.now() > expiresAt) {
      throw new Error('Canva access token expired — refresh before calling');
    }
  }

  // Block sensitive operations based on environment
  static validateEnvironment(operation: string): void {
    const blocked = ['deleteAsset', 'deleteFolder'];
    if (process.env.NODE_ENV !== 'production' && blocked.includes(operation)) {
      // Allow in dev for testing
      return;
    }
    // In production, require explicit confirmation
    if (process.env.NODE_ENV === 'production' && blocked.includes(operation)) {
      console.warn(`[guardrail] Destructive operation: ${operation}`);
    }
  }

  // Rate limit self-check — don't send if we know we'll get 429
  static checkRateLimit(
    endpoint: string,
    tracker: Map<string, { count: number; resetAt: number }>
  ): void {
    const window = tracker.get(endpoint);
    if (window && window.count <= 0 && Date.now() < window.resetAt) {
      const waitMs = window.resetAt - Date.now();
      throw new Error(`Rate limit exhausted for ${endpoint}. Wait ${(waitMs / 1000).toFixed(0)}s`);
    }
  }
}

Output

Guardrails produce an allow/deny decision, policy version, opaque subject/scope identifiers, and a redacted audit result. They never convert a missing policy or failed validation into an implicit allow.

Examples

Before publishing an export, check the tenant, asset rights, approved destination, and retention policy against the reviewed allowlist. If anything is unknown, deny the request and require owner approval rather than bypassing the guardrail in CI or runtime.

Error Handling

Issue Cause Solution
ESLint rule not firing Wrong config path Check plugin registration
Pre-commit skipped --no-verify used Enforce in CI
False positive on "OCA" String matches pattern Narrow regex or add allowlist
Guardrail blocks valid op Too strict Add environment-based exceptions

Resources

Next Steps

For architecture blueprints, see canva-architecture-variants.