Observed since 20 Oct, 23:18 UTC
No superseded metadata states collected. Unchanged observations do not create new revisions.
Tamper-proof audit trails for AI agents.
Tamper-proof audit trails for AI agents. Every action is cryptographically chained — tamper with one record and the entire chain breaks.
pip install agentseal-sdk
from agentseal import Seal
seal = Seal(api_key="as_sk_...")
# Record an action
seal.record(
agent="my-bot",
action="email:send",
params={"to": "user@example.com"},
reasoning="User requested password reset",
)
# Query the audit trail
entries = seal.query(agent="my-bot")
# Verify chain integrity
result = seal.verify(agent="my-bot")
print(result) # {"valid": True, "entries_verified": 1}
Seal(api_key, base_url=DEFAULT)Create a client. Supports with Seal(...) as seal: context manager.
seal.record(agent, action, params, reasoning, authorized_by)Record an agent action to the tamper-proof chain. Returns {id, sequence, entry_hash, parent_hash}.
seal.query(agent, action_type, since, until, limit)Query recorded entries with optional filters. Returns a list of entry dicts.
seal.verify(agent)Verify hash chain integrity. Returns {"valid": True, "entries_verified": N} or {"valid": False, "broken_at_sequence": N, "reason": "..."}.
@seal.track(agent, action)Decorator to auto-record function calls.
@seal.track(agent="my-bot")
def send_email(to, subject, body):
... # function runs, then the call is recorded automatically
Sign up at agentseal.io
MIT
Download counts and popularity will appear here as the collector receives them.
| Version | Published | Direct dependencies | Open version |
|---|---|---|---|
| 0.1.0 | 8 Apr 2026 | 1 |
This describes our collected data; it is not a security assessment.
Observed since 20 Oct, 23:18 UTC
No superseded metadata states collected. Unchanged observations do not create new revisions.
package:e062f08d5f9f551c937778a74011cb7979dc39dfaada0b16f1a844a3fb1a2e20