Prompt secret scanner
Check a prompt or assembled context for credentials, tokens, connection strings and personal data before it leaves the building.
Prompts get assembled from context nobody reviewed — a ticket body, a log excerpt, a database row. This looks for the shapes that should not be leaving the building inside one.
Scanning the prompt is the last line, not the first. The durable fix is that the context assembly step never picks up the credential in the first place — which is a data-flow problem in the layer beneath the model, not a prompt problem.
Pattern matching only, and it will miss anything shaped unusually. Card numbers are Luhn-checked so a 16-digit order number is not reported as one. Matches are masked on screen. Nothing is transmitted — which matters more here than on any other tool, since the input is by definition the thing you did not want sent anywhere.
Other tools
- EDI viewer & validator (X12 + EDIFACT)
- X.509 certificate decoder
- JSON ↔ XML converter
- CSV ↔ JSON converter
- OpenAPI linter
- MCP tool-definition linter
- OpenAPI → agent tools
- Agent blast-radius calculator
- Agent cost & context calculator
- Agent-readiness scorecard
- iPaaS readiness assessment
- Modernisation roadmap generator
- Migration risk register
- Cutover runbook builder
- Interface catalogue builder
- Migration wave planner
- Function Point estimator
- Platform migration assessor
- Fixed-width flat file parser
- webMethods flow reader
- SAP IDoc viewer
- SWIFT MT & ISO 20022 viewer
- HL7 v2 message viewer
- XSLT tester
- XPath tester
- CI/CD workflow security linter
- Regex tester & ReDoS checker
- Webhook signature verifier
- SLA & error budget calculator
- Retry & backoff simulator
- JSONPath tester
- JSON Schema generator
- Encoding & mojibake fixer
- JSON formatter & validator
- JSON diff
- Epoch & timestamp converter
- Base64 encoder & decoder
- UUID & ULID generator
- JWT decoder
- Cron expression explainer
- Interface math calculator
- Post-quantum migration matrix
- Post-quantum exposure calculator
These exist because the underlying problem is real. If the numbers you just put in look uncomfortable, that is usually worth a conversation.