Repository navigation
Breach Exposure Check AgentFlow
Use Have I Been Pwned (HIBP) breach metadata to help an organization or account owner assess exposure. Run checks only for an account owner, a verified domain administrator, or an assessment expressly authorized in writing.
HIBP reports whether an address appears in known breach records and which data classes were involved. Its account and paste endpoints do not return stolen passwords, hashes, or dump contents. A match does not prove that an account is currently compromised or that a password is reusable.
- Check only addresses covered by documented consent or an approved organizational scope.
- Domain searches require HIBP domain verification and an API key with the appropriate access.
- Treat queried addresses and results as sensitive personal data. Minimize what enters prompts, logs, and reports; follow retention policy.
- Never test credentials, build a password-spraying list, or ask the model to infer passwords from breach metadata.
- Store API keys in the platformโs secret store. Do not put secrets in URLs, prompts, flow variables, or exported flow definitions.
- Review HIBPโs current API documentation for authentication, access, response fields, and rate limits.
| Variable | Description |
|---|---|
email_address |
One authorized email address, URL-encoded by the connector |
breach_metadata |
HIBP breach metadata response |
paste_metadata |
Optional HIBP paste metadata response |
summary |
Minimized analyst summary and owner-focused next steps |
Do not use one input field interchangeably for an email address and a domain. For verified-domain monitoring, configure the separate breacheddomain/{domain} operation and verify the domain through HIBP first.
Configure the HTTP/API block to call:
GET https://haveibeenpwned.com/api/v3/breachedaccount/{URL_ENCODED_EMAIL}
Use the hibp-api-key header and a descriptive user-agent header, with values supplied by the approved secret manager or connector settings. Do not paste a real key into this document. A no-match response may be HTTP 404; handle that as โno matching breach returned,โ not as proof that the address is safe.
If the authorized workflow needs this check and the API subscription permits it, call:
GET https://haveibeenpwned.com/api/v3/pasteaccount/{URL_ENCODED_EMAIL}
This returns paste metadata, not paste contents or credentials. Omit this step if the result is not needed for the stated purpose.
Give the model only the fields needed for the report, such as breach name, breach date, and listed data classes. Exclude account addresses where practical, and never include credentials or raw breach material.
Prompt:
Summarize the supplied HIBP metadata for the authorized account owner.
Report only the breach name, breach date, and data classes explicitly present.
Do not infer passwords, password reuse, account takeover, identity, or user behavior.
State that a result is historical exposure metadata and does not establish current compromise.
If the input is empty or malformed, say that the lookup result is unavailable.
Recommend owner-directed actions: use a unique password, enable MFA or passkeys,
review account sign-in activity, and follow the organizationโs incident process
if there is independent evidence of compromise.
Lookup: completed for an authorized account (address omitted)
Result: one matching breach record
Breach: Example Service (date reported by HIBP)
Data classes: email addresses, usernames
Interpretation: historical exposure metadata; no password or current account access was assessed.
Next steps: account owner should use a unique password and MFA/passkeys; investigate further only if separate evidence warrants it.
- Verify the account or domain is in scope before querying.
- Record the authorization basis, lookup date, source, and minimal result needed for the report.
- A missing result is not proof that an account was never exposed; HIBP coverage is not exhaustive.
- Do not mandate an organization-wide password reset solely from breach metadata. Follow the incident response process and assess current risk using approved identity-provider signals.
- If accidental sensitive data enters the flow, stop processing, restrict access, and follow the organizationโs data-handling procedure.
- EmailOSINT โ consent-based account security checks
- CompanyOSINT โ organization-level public information
๐ Source: AI/AnythingLLM/flow-breacher_check.md | โ๏ธ Suggest an edit | โฌ
๏ธ Back to Artificial Intelligence
- ๐บ๏ธ Browse by Topic
- ๐ All Repository Files
- ๐งญ Learning Paths
- ๐ Master Guide
- ๐ค Glossary
- โ๏ธ Legal & Disclaimer
๐ Foundations & Learning
๐ก๏ธ Defense & Investigations
- Incident Response
- Open Source Intelligence (OSINT)
- Reverse Engineering
- Operations Security (OPSEC)
- Compliance
โ๏ธ Platforms & Infrastructure
๐ก Hardware, Wireless & Field Projects
- PNWC Cyberdeck
- Hardware Testing
- Hardware Hacking
- Software-Defined Radio
- Wardriving
- uConsole
- Space Security
๐งฐ Assessment & Field Reference
๐ฆ Downloads & Project Resources
๐๏ธ Additional Topics