X402 RISK ORACLE FOR AUTONOMOUS SOFTWARE ACTIONS

Proof before permission.

Sigillum gives an AI agent, CLI user, or GitHub workflow a neutral, paid risk receipt before a code change, dependency install, deploy action, or similar software operation continues.

402 quotedUSDC paidunits inspectedscore returnedreceipt signed

AI-GENERATED DIFF

+ API_KEY="sk_live_..."
+ eval(plugin.payload)
+ "Deploy succesful"

HTTP 402

Payment Required0.000043 USDC via x402
Sigillum ReceiptBLOCK
82

Sigillum Score

PAID
0.000043 USDC
INSPECTED UNITS
431
SEAL
Verified by Sigillum
TRUST GAP

NEUTRAL REVIEW

The agent that writes the code should not be the only one approving it.

AI coding tools can generate production changes in seconds. Sigillum creates an external checkpoint at the risky moment before merge, deploy, install, or publish.

SELF-REVIEW RISK

Same system wrote and approved the change.

External receipt required before action continues.

01Agent wrote change
02Self-review is weak
03External receipt required
PROOF LOOP

QUOTE. PAY. INSPECT. ACT.

A paid proof loop for risky software actions.

The landing page shows the product contract without running the live flow here. The full interactive quote, 402 payment gate, receipt, and decision journey belongs on the dashboard.

01

Submit diff

Changed surface only

02

HTTP 402

0.000043 USDC

03

Inspect

431 units checked

04

Receipt

sig_01J7K9...

05

Decision

Merge blocked

SIGNED PROOF OBJECT

The receipt is the product.

A paid proof object your agent can act on, with score, findings, payment, inspected units, and recommendation in one machine-readable artifact.

SIGILLUM RECEIPT

Verified by Sigillum

BLOCK
82

Sigillum Score

431

Inspected Units

0.000043

USDC Paid

CRITICAL FINDINGPossible API key added in .env.example

Patch recommendation: remove the exposed key and rotate the credential before merge.

RECEIPT ID sig_01J7K9X4M6A2

{
  "receipt_id": "sig_01J7K9X4M6A2",
  "seal": "Verified by Sigillum",
  "score": 82,
  "recommendation": "block",
  "paid_amount_usdc": "0.000043",
  "inspected_units": {
    "ast_nodes": 312,
    "changed_lines": 74,
    "dependency_changes": 2,
    "config_mutations": 1,
    "strings": 42
  },
  "findings": [
    {
      "severity": "critical",
      "category": "secret_exposure",
      "message": "Possible API key added in .env.example",
      "file": ".env.example",
      "line": 4
    }
  ],
  "patch_recommendation": "Remove the exposed key and rotate the credential before merge."
}
INSPECTION BOARD

RISK SIGNALS

Small checks before large mistakes.

Sigillum checks the changed surface for risk signals. It does not claim guaranteed security; it returns bounded evidence and a recommendation.

critical

Secrets

Possible API key added

warn

Dependency risk

2 package changes

warn

Config mutation

1 deployment setting

blocked

Dangerous APIs

eval-like path found

pass

Prompt surfaces

No prompt override

pass

Syntax structure

Parsed successfully

needs patch

Copy issues

User-facing typo

X402 PRICING

No subscription. Pay per inspected unit.

Some checks are too small for monthly billing. Sigillum prices each verification by changed surface area, so agents buy proof only when they need it.

Changed lines74$0.000074
AST nodes312$0.000156
Dependency changes2$0.000040
Config mutations1$0.000010
Total paid0.000043 USDC

AGENT DECISION

The receipt controls the next action.

Risk score82
RecommendationBLOCK
Policy matchedblock_on.secret_exposure
Agent actionstop merge
Next actionremove secret and regenerate patch
DEVELOPER CHECKPOINT

AGENT-CALLABLE API

One checkpoint before risky action.

Sigillum is designed for agents to call before they act. The dashboard hosts the live receipt flow, while the landing page keeps the contract visible.

POST /quote
-> quote_01J...
-> 0.000043 USDC

POST /inspect
-> HTTP 402 Payment Required
-> Payment confirmed
-> Receipt generated

RESULT SUMMARY

Machine-readable recommendation

{
  "recommendation": "block",
  "score": 82,
  "receipt_id": "sig_01J..."
}