Gamma Data HandlingSAFE

Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: gamma-data-handling
description: 'Handle data privacy, retention, and compliance for Gamma integrations.

  Use when implementing GDPR compliance, data retention policies,

  or managing user data within Gamma workflows.

  Trigger with phrases like "gamma data", "gamma privacy",

  "gamma GDPR", "gamma data retention", "gamma compliance".

  '
allowed-tools: Read, Write, Edit
version: 1.13.0
license: MIT
author: Jeremy Longshore <[email protected]>
tags:
- saas
- gamma
- workflow
- compliance
compatibility: Designed for Claude Code
---
# Gamma Data Handling

## Output

Record source reference, approved field/destination policy, retention rule, aggregate validation result, owner, and exception state. Keep presentation content, viewer data, and credentials outside the receipt.

## Examples

Process a fictional staging deck through a field allowlist, confirm only approved metadata reaches the test destination, and remove the fixture according to the retention policy.

## Overview

Data handling, privacy controls, and compliance for Gamma API integrations. Gamma processes user-submitted content through AI to generate presentations -- understand what data flows where and how to handle PII, retention, and GDPR requirements.

## Prerequisites

- Understanding of data privacy regulations (GDPR, CCPA)
- Completed `gamma-install-auth` setup
- Data classification policies defined

## Data Flow Map

```
User Input (content, prompts)
     │
     ▼
┌──────────────┐
│  Your App    │  ← PII may be in content (names, company data)
│  (API key)   │
└──────┬───────┘
       │ POST /v1.0/generations
       ▼
┌──────────────┐
│  Gamma API   │  ← Content processed by AI
│  (gamma.app) │  ← Images generated
└──────┬───────┘
       │ gammaUrl + exportUrl
       ▼
┌──────────────┐
│  Generated   │  ← Presentation stored in Gamma workspace
│  Content     │  ← Export files (PDF/PPTX/PNG) temporary
└──────────────┘
```

## Data Classification

| Data Type | Classification | Where Stored | Retention |
|-----------|---------------|--------------|-----------|
| API key | Secret | Your env vars | Active use only |
| Content/prompts | May contain PII | Gamma servers (during generation) | Gamma's policy |
| Generated gammas | User data | Gamma workspace | User-controlled |
| Export files (PDF/PPTX) | User data | Temporary URLs | Download promptly, URLs expire |
| User prompts in logs | PII risk | Your infrastructure | Your policy (sanitize!) |
| Credit usage | Billing data | Gamma | Per Gamma ToS |

## Instructions

### Step 1: Sanitize Content Before Sending

```typescript
// src/gamma/sanitize.ts
// Remove PII from content before sending to Gamma if not needed

interface SanitizeOptions {
  removeEmails: boolean;
  removePhones: boolean;
  maskNames: boolean;
}

function sanitizeContent(content: string, opts: SanitizeOptions): string {
  let sanitized = content;

  if (opts.removeEmails) {
    sanitized = sanitized.replace(/[\w.-]+@[\w.-]+\.\w+/g, "[email]");
  }
  if (opts.removePhones) {
    sanitized = sanitized.replace(/\+?[\d\s()-]{10,}/g, "[phone]");
  }
  if (opts.maskNames) {
    // Only mask if you have a list of known names
    // Generic regex would be too aggressive
  }

  return sanitized;
}

// Usage: sanitize before generation
const safeContent = sanitizeContent(userContent, {
  removeEmails: true,
  removePhones: true,
  maskNames: false,
});

await gamma.generate({
  content: safeContent,
  outputFormat: "presentation",
});
```

### Step 2: Sanitize Logs

```typescript
// src/gamma/logging.ts
// Never log raw content or API keys

function logGeneration(request: any, result: any) {
  console.log(JSON.stringify({
    event: "gamma_generation",
    timestamp: new Date().toISOString(),
    generationId: result.generationId,
    outputFormat: request.outputFormat,
    contentLength: request.content?.length,
    // NEVER log: content (may have PII), apiKey
    status: result.status,
    creditsUsed: result.creditsUsed,
  }));
}
```

### Step 3: Export File Handling

```typescript
// src/gamma/exports.ts
// Export URLs are temporary — download and store securely

import { writeFile } from "node:fs/promises";
import { S3Client, PutObjectCommand } from "@aws-sdk/client-s3";

async function archiveExport(
  exportUrl: string,
  metadata: { generationId: string; userId: string }
) {
  // Download immediately — URLs expire
  const res = await fetch(exportUrl);
  if (!res.ok) throw new Error(`Export download failed: ${res.status}`);
  const buffer = Buffer.from(await res.arrayBuffer());

  // Store with encryption
  const s3 = new S3Client({ region: "us-east-1" });
  const key = `gamma-exports/${metadata.userId}/${metadata.generationId}.pdf`;

  await s3.send(new PutObjectCommand({
    Bucket: process.env.EXPORTS_BUCKET!,
    Key: key,
    Body: buffer,
    ContentType: "application/pdf",
    ServerSideEncryption: "aws:kms",
    Metadata: {
      generationId: metadata.generationId,
      archivedAt: new Date().toISOString(),
    },
  }));

  console.log(`Archived: s3://${process.env.EXPORTS_BUCKET}/${key}`);
}
```

### Step 4: Data Retention Policy

```typescript
// src/gamma/retention.ts
interface RetentionPolicy {
  exportMaxDays: number;     // Delete local export copies
  logRetentionDays: number;  // Anonymize generation logs
  promptRetentionDays: number; // Delete stored prompts
}

const POLICY: RetentionPolicy = {
  exportMaxDays: 90,       // Keep exports 90 days
  logRetentionDays: 30,    // Anonymize logs after 30 days
  promptRetentionDays: 7,  // Delete prompts after 7 days
};

async function enforceRetention() {
  const cutoff = new Date();

  // Delete old exports from S3
  cutoff.setDate(cutoff.getDate() - POLICY.exportMaxDays);
  await deleteOldExports(cutoff);

  // Anonymize old logs
  cutoff.setDate(cutoff.getDate() + POLICY.exportMaxDays - POLICY.logRetentionDays);
  await anonymizeLogs(cutoff);

  // Delete stored prompts
  cutoff.setDate(cutoff.getDate() + POLICY.logRetenti

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

What the source does

Filesystem

none-observed

Network

none-observed

Shell

none-observed

Dependencies

pinned

Secrets in source

none-found

Findings (0)

No findings outside the package's declared scope.

Gates applied: no_behavioural_pass.