Verifiable AI Agent Memory_Cryptographic audit trail for CrewAI agents.
Every agent decision, task execution, and crew result — signed, hash-chained, tamper-proof.
pip install piqrypt-crewaiCrewAI agents make decisions autonomously. When something goes wrong — or needs to be proven right — your logs are editable. PiQrypt fixes that.
from piqrypt_crewai import AuditedAgent, AuditedCrew, stamp_task
# Drop-in replacement for Agent
researcher = AuditedAgent(
role="Researcher",
goal="Find accurate information on AI compliance",
backstory="Expert in regulatory frameworks",
identity_file="researcher.json" # piqrypt identity
)
writer = AuditedAgent(
role="Writer",
goal="Write clear compliance reports",
backstory="Technical writer",
identity_file="writer.json"
)
# Drop-in replacement for Crew
crew = AuditedCrew(
agents=[researcher, writer],
tasks=[research_task, write_task],
identity_file="crew-coordinator.json"
)
result = crew.kickoff(inputs={"topic": "EU AI Act compliance"})
# Export tamper-proof audit trail
crew.export_audit("eu-ai-act-audit.json")
# $ piqrypt verify eu-ai-act-audit.jsonfrom piqrypt_crewai import stamp_task
@stamp_task("research", identity_file="my-agent.json")
def research(topic: str) -> str:
return your_research_logic(topic)
@stamp_task("write_report", identity_file="my-agent.json")
def write_report(research: str) -> str:
return your_writing_logic(research)| Event | When |
|---|---|
agent_initialized |
Agent creation |
task_start |
Before task execution |
task_complete |
After task execution (with result hash) |
crew_kickoff |
Before crew runs |
crew_complete |
After crew finishes (with result hash) |
task_executed |
When using @stamp_task decorator |
All events are Ed25519-signed, SHA-256 hash-chained, stored locally.
No content is stored — only SHA-256 hashes of inputs and outputs.
piqrypt verify crewai-audit.json
# ✅ Chain integrity verified — 12 events, 0 forks| Use case | AISS profile |
|---|---|
| Development / PoC | AISS-1 (Free, included) |
| Non-critical production | AISS-1 (Free) |
| Regulated production | AISS-2 (Pro — pip install piqrypt[aiss2]) |
- PiQrypt core: github.com/piqrypt/piqrypt
- Integration guide: INTEGRATION.md
- Issues: github.com/piqrypt/piqrypt/issues