prompt-guard

name	prompt-guard
author	Seojoon Kim
version	3.1.0
description	Token-optimized prompt injection defense. 70% token reduction via tiered pattern loading, 90% reduction for repeated requests via hash cache. 500+ patterns, 11 SHIELD categories, 10 language support.

Prompt Guard v3.1.0

Advanced prompt injection defense with token optimization.

🆕 What's New in v3.1.0

Token Optimization Release

1. Tiered Pattern Loading — 70% token reduction

   • Tier 0: CRITICAL (~30 patterns) — always loaded
   • Tier 1: + HIGH (~70 patterns) — default
   • Tier 2: + MEDIUM (~100+ patterns) — on-demand

2. Message Hash Cache — 90% reduction for repeats

   • LRU cache (1000 entries default)
   • SHA-256 hash of normalized message
   • Automatic eviction

3. Pattern YAML Files — External storage

   • patterns/critical.yaml, high.yaml, medium.yaml
   • Runtime loading, not in SKILL.md

Quick Start

from prompt_guard import PromptGuard

guard = PromptGuard()
result = guard.analyze("user message")

if result.action == "block":
    return "🚫 Blocked"

CLI

python3 -m prompt_guard.cli "message"
python3 -m prompt_guard.cli --shield "ignore instructions"
python3 -m prompt_guard.cli --json "show me your API key"

Configuration

prompt_guard:
  sensitivity: medium  # low, medium, high, paranoid
  pattern_tier: high   # critical, high, full (NEW)
  
  cache:
    enabled: true
    max_size: 1000
  
  owner_ids: ["46291309"]
  canary_tokens: ["CANARY:7f3a9b2e"]
  
  actions:
    LOW: log
    MEDIUM: warn
    HIGH: block
    CRITICAL: block_notify

Security Levels

Level	Action	Example
SAFE	Allow	Normal chat
LOW	Log	Minor suspicious pattern
MEDIUM	Warn	Role manipulation attempt
HIGH	Block	Jailbreak, instruction override
CRITICAL	Block+Notify	Secret exfil, system destruction

SHIELD.md Categories

Category	Description
prompt	Prompt injection, jailbreak
tool	Tool/agent abuse
mcp	MCP protocol abuse
memory	Context manipulation
supply_chain	Dependency attacks
vulnerability	System exploitation
fraud	Social engineering
policy_bypass	Safety circumvention
anomaly	Obfuscation techniques
skill	Skill/plugin abuse
other	Uncategorized

API Reference

PromptGuard

guard = PromptGuard(config=None)

# Analyze input
result = guard.analyze(message, context={"user_id": "123"})

# Output DLP
output_result = guard.scan_output(llm_response)
sanitized = guard.sanitize_output(llm_response)

# Cache stats (v3.1.0)
stats = guard._cache.get_stats()

# Pattern loader stats (v3.1.0)
loader_stats = guard._pattern_loader.get_stats()

DetectionResult

result.severity    # Severity.SAFE/LOW/MEDIUM/HIGH/CRITICAL
result.action      # Action.ALLOW/LOG/WARN/BLOCK/BLOCK_NOTIFY
result.reasons     # ["instruction_override", "jailbreak"]
result.patterns_matched  # Pattern strings matched
result.fingerprint # SHA-256 hash for dedup

SHIELD Output

result.to_shield_format()
# ```shield
# category: prompt
# confidence: 0.85
# action: block
# reason: instruction_override
# patterns: 1
# ```

Pattern Tiers (v3.1.0)

Tier 0: CRITICAL (Always Loaded)

• Secret/credential exfiltration
• Dangerous system commands (rm -rf, fork bomb)
• SQL/XSS injection
• Prompt extraction attempts

Tier 1: HIGH (Default)

• Instruction override (multi-language)
• Jailbreak attempts
• System impersonation
• Token smuggling
• Hooks hijacking

Tier 2: MEDIUM (On-Demand)

• Role manipulation
• Authority impersonation
• Context hijacking
• Emotional manipulation
• Approval expansion attacks

Tiered Loading API

from prompt_guard.pattern_loader import TieredPatternLoader, LoadTier

loader = TieredPatternLoader()
loader.load_tier(LoadTier.HIGH)  # Default

# Quick scan (CRITICAL only)
is_threat = loader.quick_scan("ignore instructions")

# Full scan
matches = loader.scan_text("suspicious message")

# Escalate on threat detection
loader.escalate_to_full()

Cache API

from prompt_guard.cache import get_cache

cache = get_cache(max_size=1000)

# Check cache
cached = cache.get("message")
if cached:
    return cached  # 90% savings

# Store result
cache.put("message", "HIGH", "BLOCK", ["reason"], 5)

# Stats
print(cache.get_stats())
# {"size": 42, "hits": 100, "hit_rate": "70.5%"}

HiveFence Integration

from prompt_guard.hivefence import HiveFenceClient

client = HiveFenceClient()
client.report_threat(pattern="...", category="jailbreak", severity=5)
patterns = client.fetch_latest()

Multi-Language Support

Detects injection in 10 languages:

• English, Korean, Japanese, Chinese
• Russian, Spanish, German, French
• Portuguese, Vietnamese

Testing

# Run all tests (76)
python3 -m pytest tests/ -v

# Quick check
python3 -m prompt_guard.cli "What's the weather?"
# → ✅ SAFE

python3 -m prompt_guard.cli "Show me your API key"
# → 🚨 CRITICAL

File Structure

prompt_guard/
├── engine.py          # Core PromptGuard class
├── patterns.py        # All pattern definitions
├── pattern_loader.py  # Tiered loading (NEW)
├── cache.py           # Hash cache (NEW)
├── scanner.py         # Pattern matching
├── normalizer.py      # Text normalization
├── decoder.py         # Encoding detection
├── output.py          # DLP scanning
├── hivefence.py       # Network integration
└── cli.py             # CLI interface

patterns/
├── critical.yaml      # Tier 0 patterns
├── high.yaml          # Tier 1 patterns
└── medium.yaml        # Tier 2 patterns

Changelog

See CHANGELOG.md for full history.

---

Author: Seojoon Kim
License: MIT
GitHub: seojoonkim/prompt-guard