notaryos

Seal AI agent actions with Ed25519 cryptographic receipts. Verify what your agent did and prove what it chose not to do.

Safety Notice

This listing is from the official public ClawHub registry. Review SKILL.md and referenced scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "notaryos" with this command: npx skills add hellothere012/notaryos

NotaryOS — Cryptographic Receipts for Agent Actions

Seal your agent's actions with Ed25519 signatures. Issue tamper-evident receipts, verify them publicly, and maintain an auditable chain of every decision.

License

BSL-1.1 (Business Source License). See https://github.com/hellothere012/notaryos/blob/main/LICENSE

Trust Statement

By using this skill, action metadata (action type, timestamps, and a SHA-256 hash of the payload) is sent to api.agenttownsquare.com via HTTPS. Raw payload retention depends on your tier — see the Data Flow section below. Verification is free and requires no account. Full privacy policy: https://notaryos.org/privacy

Data Flow

The SDK sends your payload to the NotaryOS API via HTTPS POST. The server hashes the payload with SHA-256, signs the hash with Ed25519, and returns a receipt.

TierPayload TransmittedRaw Payload RetainedHash StoredSignature Stored
Demo (no key)YesNoYesYes
FreeYesMetadata onlyYesYes
ProYesConfigurableYesYes
EnterpriseYesZero retentionYesYes

The included sanitize.py module strips fields matching known sensitive patterns before transmission. Use it before every seal() call when handling user data.

External Endpoints

URLMethodData SentPurpose
api.agenttownsquare.com/v1/notary/issuePOSTaction_type, payload JSONIssue signed receipt
api.agenttownsquare.com/v1/notary/verifyPOSTreceipt JSONVerify signature
api.agenttownsquare.com/v1/notary/statusGETNoneHealth check
api.agenttownsquare.com/v1/notary/r/{hash}GETNoneReceipt lookup
api.agenttownsquare.com/v1/notary/public-keyGETNoneEd25519 public key

No other endpoints are contacted. No telemetry, analytics, or tracking.

Setup

pip install notaryos

No API key required. The SDK auto-injects a free demo key (10 req/min) when NOTARY_API_KEY is not set. For production rates, get a key at https://notaryos.org/sign-up and set NOTARY_API_KEY in your environment or OpenClaw config.

from notaryos import NotaryClient

notary = NotaryClient()  # works immediately — uses demo key if NOTARY_API_KEY is not set

Seal an Action

from notaryos import NotaryClient
from sanitize import sanitize_payload

notary = NotaryClient()

receipt = notary.seal(
    "file.created",
    sanitize_payload({
        "path": "/src/main.py",
        "lines_added": 42,
        "branch": "feature/auth"
    })
)

print(receipt.receipt_hash)
print(receipt.signature)

What to Seal

Default (always safe)

Action TypeWhen to Seal
file.createdCreated or modified a file
file.deletedDeleted a file
command.executedRan a shell command
config.changedModified system configuration

Extended (sanitize payload first)

Action TypeWhen to Seal
email.sentSent an email (strip body, keep subject)
api.calledMade an external API call (strip auth headers)
data.accessedAccessed sensitive data (log access, not content)
message.sentSent a message (strip body if private)

Always run sanitize_payload() on extended actions before sealing.

Payload Guidelines

Include: File paths, counts, timestamps, branch names, public identifiers, action summaries.

Exclude: Authentication credentials, financial numbers, government IDs, message bodies, file contents, health information. The sanitize_payload() helper handles this automatically.

Verify a Receipt

from notaryos import verify_receipt

is_valid = verify_receipt(receipt.to_dict())  # True or False, no auth needed

Lookup by Hash

notary = NotaryClient()
result = notary.lookup("e1d66b0bdf3f8a7e...")

if result["found"] and result["verification"]["valid"]:
    print("Receipt is authentic and untampered")

Counterfactual Receipts

Record when your agent chose NOT to act:

receipt = notary.seal("trade.declined", {
    "reason": "risk_threshold_exceeded",
    "action_considered": "trade.execute",
    "decision": "blocked"
})

Receipt Chaining

r1 = notary.seal("file.read", {"file": "report.pdf"})
r2 = notary.seal("summary.generated", {
    "source": "report.pdf",
    "length": 500
}, previous_receipt_hash=r1.receipt_hash)

Error Handling

from notaryos import AuthenticationError, RateLimitError, ValidationError

try:
    receipt = notary.seal("action", {"key": "value"})
except RateLimitError:
    pass  # demo: 10 req/min, upgrade at notaryos.org
except AuthenticationError:
    pass  # invalid key
except ValidationError:
    pass  # bad request

Dependencies

Key Points

  • NOTARY_API_KEY is optional — a demo key is auto-injected when not set (10 req/min)
  • Set NOTARY_API_KEY for production rates (get a key at https://notaryos.org/sign-up)
  • Both sanitize.py and the notaryos SDK use only the Python standard library (zero third-party deps)
  • Payloads transmitted via HTTPS to api.agenttownsquare.com
  • Use sanitize_payload() to strip sensitive fields before sealing
  • Verification is free and public — no API key needed
  • Ed25519 signatures (same scheme as SSH and TLS)

Links

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

Security

Agentshield Audit

Trust Infrastructure for AI Agents - Like SSL/TLS for agent-to-agent communication. 77 security tests, cryptographic certificates, and Trust Handshake Protoc...

Registry SourceRecently Updated
1.2K0Profile unavailable
Security

Receipts Guard

ERC-8004 identity, x402 payments, and arbitration protocol for autonomous agent commerce. The three rails for the machine economy.

Registry SourceRecently Updated
2.1K1Profile unavailable
Security

AxonFlow Governance Policies

Govern OpenClaw with AxonFlow — block dangerous commands, detect PII, prevent data exfiltration, protect agent config files, explain policy decisions, grant...

Registry SourceRecently Updated
Security

Vorim AI — Agent Identity & Trust

AI agent identity, permissions, trust scores, and audit trails via Vorim AI. Use when: (1) performing sensitive actions that need permission checks, (2) logg...

Registry SourceRecently Updated
380Profile unavailable