privacy-scraper
Scrape web pages privately with encrypted local storage, optional PII removal, and secure memory-only retrieval - featuring explicit zero-local-data-retention mode and verifiable memory wiping.
Privacy-First Web Scraper Skill
When to use
Use when you need to harvest information from public sites while keeping raw data local (e.g., for compliance, personal privacy, or air-gapped environments).
CRITICAL PRIVACY CONTROLS
This skill provides EXPLICIT, UNMISTAKABLE controls for data retention and sharing:
🔒 ZERO LOCAL RETENTION MODE (MAXIMUM PRIVACY)
- Flag:
--store-excerpt false - Result:
- ✅ ONE file created:
vault/<timestamp>.enc(AES-256-GCM encrypted full scrape) - ❌ ZERO files created: No excerpt file, no temporary files, no cache files
- ✅ User choice: User explicitly decides whether to keep/delete the encrypted vault file
- 📝 Use case: Highly sensitive data where ANY local trace is unacceptable
- ✅ ONE file created:
📤 STANDARD PRIVACY (STANDARD USE)
- Flag:
--store-excerpt true(default) +--remove-pii true(default) - Result:
- ✅ TWO files created:
vault/<timestamp>.enc(encrypted full scrape)excerpt.md(short summary with PII removed)
- ⚠️ Sharing caveat: The
excerpt.mdfile has PII stripped, but pattern-based removal is best-effort. Review for residual identifiers or confidential content before sharing. - ❌ Never share: The encrypted
.encfile without explicit authorization - 📝 Use case: Standard privacy-conscious sharing where excerpt distribution is needed
- ✅ TWO files created:
🔬 FORENSIC MODE (COMPLETE PRESERVATION)
- Flag:
--store-excerpt true+--remove-pii false - Result:
- ✅ TWO files created: Encrypted vault + FULL excerpt (no PII removed)
- 📝 Use case: Legal, audit, or forensic requirements needing complete records
What it does
- 🌐 ISOLATED BROWSING: Network access limited to user-provided allow-lists (default: BLOCK ALL)
- 📄 CONTENT EXTRACTION: Visible text via DOM inspection (navigation/ads filtered)
- 🔐 STRONG ENCRYPTION: AES-256-GCM with cryptographically secure random nonce per operation
- 🛡️ OPTIONAL PII REMOVAL: Configurable stripping of sensitive identifiers BEFORE any storage
- 💾 SECURE VAULT: Timestamp-named encrypted files for organization/retention
- 📋 CONDITIONAL EXCERPT: Generated ONLY when
store_excerpt=true(user-controlled) - 🔓 RETRIEVAL: Decrypted data loads into RAM during active use; not intentionally written to persistent storage
- ⚠️ ENVIRONMENTAL LIMITATIONS: Data may persist in swap space, crash dumps, or stdout capture buffers depending on the host. These are system-level behaviors that application code cannot eliminate.
- 📊 AUDIT TRAIL: SHA-256 hashes for integrity verification of stored files
- ⏱️ RESOURCE CLEANUP: Timeouts and proper browser context termination
Typical workflow
STEP 1: SELECT YOUR PRIVACY LEVEL
Choose based on data sensitivity and sharing requirements:
- [MAXIMUM PRIVACY]
--store-excerpt false→ Encrypted vault only, no local traces - [STANDARD PRIVACY]
--store-excerpt true+--remove-pii true→ Vault + excerpt (review before sharing) - [FORENSIC MODE]
--store-excerpt true+--remove-pii false→ Complete records retained
STEP 2: EXECUTE THE SCRAPE
privacy-scraper scrape \
--url "https://target-site.com/page" \
[--allow-domains "allowed-site.com,cdn.allowed.com"] \
[--timeout-seconds 30] \
[--remove-pii true|false] \
[--store-excerpt true|false] \
[--output-dir "/custom/vault/location"]
STEP 3: INTERPRET RESULTS
ALWAYS CREATED (regardless of other options):
- 🔒
vault/<timestamp>.enc- AES-256-GCM encrypted full scrape (user controls retention/deletion)
CONDITIONALLY CREATED (only when store_excerpt=true):
- 📄
excerpt.md- Short summary (PII removed ifremove_pii=true)
STEP 4: SHARE RESPONSIBLY (IF APPLICABLE)
- If
store_excerpt=trueANDremove_pii=true:- ⚠️ REVIEW BEFORE SHARING:
excerpt.mdhas PII stripped but pattern-based removal is not infallible. Inspect for residual identifiers or confidential content. - 🚫 NEVER SHARE:
vault/<timestamp>.encwithout explicit authorization
- ⚠️ REVIEW BEFORE SHARING:
- If
store_excerpt=false:- ✅ MAXIMUM PRIVACY: No excerpt files exist to share
STEP 5: SECURE RETRIEVAL (WHEN NEEDED)
privacy-scraper retrieve --id "20261020_143022_abc123" [--verify-wipe]
RETRIEVAL BEHAVIOR:
- 💾 Disk interaction: Reads the encrypted
.encfile; decrypted data loads into RAM. Not intentionally written to persistent storage. - ⚠️ Environmental limitations: Decrypted content may persist in swap, stdout capture, or crash dumps depending on the host system — these are not under the tool's control.
- 🧠 RAM residence: Data exists in memory while the caller is actively using it.
- 🔐 Integrity check: Automatic SHA-256 hash verification against original vault file
- 📝 Output: Decrypted text returned to caller (cleared from memory after use)
STEP 6: VERIFY INTEGRITY (ANYTIME)
Check SHA-256 hash of vault/<timestamp>.enc against original record to detect tampering
Security & Privacy Guarantees
🛡️ EXTERNAL DATA TRANSMISSION
- Default: Blocks all external requests beyond allow-list
- Override: Only via explicit
--allow-domainsuser specification
💾 LOCAL DATA STORAGE BEHAVIOR
- Always created: One encrypted vault file per scrape operation
- Conditionally created: Excerpt file ONLY when user explicitly sets
store_excerpt=true - User control: Individual decides retention/deletion of encrypted vault file
- No hidden storage: Zero temporary files, caches, or artifacts created unintentionally
🔐 CRYPTOGRAPHY & INTEGRITY
- Encryption: AES-256-GCM with cryptographically secure random nonce per operation
- Integrity: SHA-256 hashing for tamper detection on stored files
- Verification: Automatic hash matching during retrieval process
🧹 MEMORY MANAGEMENT
- Technique: Secure zeroization (DoD 5220.22-M / NIST SP 800-88)
- Limitations: Swap, crash dumps, and stdout capture are host-level behaviors that cannot be eliminated by application code
Configuration Options
$PRIVACY_SCRAPER_VAULT_DIR: Custom directory for encrypted storage (default:<skill-dir>/vault)$PRIVACY_SCRAPER_KEY_FILE: File containing base64-encoded encryption key for session consistency$PRIVACY_SCRAPER_DEFAULT_TIMEOUT: Page load timeout in seconds (default: 30)$PRIVACY_SCRAPER_REMOVE_PII_DEFAULT: PII removal default (default: true)$PRIVACY_SCRAPER_LOG_LEVEL: Operational verbosity (error, warn, info, debug)$PRIVACY_SCRAPER_SECURE_WIPE_PASSES: Memory wipe overwrite passes (default: 3)
Supported Features
- Browser Controls: Isolated context with disabled extensions/plugins, network restrictions
- Content Extraction: DOM-based main content identification (filters chrome/advertising)
- Encryption Standard: AES-256-GCM - authenticated encryption with confidentiality+integrity
- Integrity Protection: SHA-256 hashing - tamper evidence for stored files
- PII Detection: Comprehensive patterns for emails, phones, SSNs, CCs, IBANs, addresses, names, IDs
- Cross-Platform: Linux/macOS/Windows with appropriate browser dependencies
Triggers
This skill activates when the user explicitly requests web scraping or vault retrieval from this tool. It does NOT activate for general privacy advice, compliance discussions, or unrelated encryption tasks.
Activation examples
- "scrape website with zero local files retained"
- "private scraping encrypted storage only no excerpt"
License
MIT — Copyright (c) 2026 z-hussein
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files, to deal in the Software without restriction. See LICENSE for details.
Related skills
- Downloads
- 385All time
- Security audit
- Pass
- Last updated
- 2d ago
- Current version
- v1.0.3
- License
- MIT-0
