Commands

scan

Run a token through a set of heuristic checks in one pass: decode it, flag common weaknesses, try weak secrets on HS tokens, and suggest matching attack payloads. Use it to triage a token before deciding which of crack, payload, or verify to run next.

Usage

jwt-hack scan <TOKEN> [OPTIONS]

▎ SCAN ───────────────────────────────────────

  Algorithm         HS256
  Type              JWT

▎ RESULTS ────────────────────────────────────

  ✓ PASS  None Algorithm         Token does not use 'none' algorithm
  ✓ PASS  Algorithm Confusion    Symmetric algorithm
  ...
  ▲ CRIT  Weak Secret            Uses weak secret: 'test'
  ◆ MED   Token Expiration       Missing 'exp', 'nbf', 'iat'
  ■ LOW   Missing Claims         Missing recommended claims: aud, iss, jti

▎ SUMMARY ────────────────────────────────────

  3 vulnerabilities found: 1 critical, 1 medium, 1 low

What it checks

For a standard JWS (3-part token):

none algorithm in use
algorithm confusion risk: asymmetric algs are flagged for follow-up, not confirmed
weak/guessable HMAC secret (HS* only; see below)
kid header present (SQL/path-injection surface)
jku / x5u headers (URL spoofing, remote JWKS)
embedded jwk header
crit, b64 (RFC 7797), zip, and typ header misuse
alg edge values and whether the signature segment is present
ECDSA psychic-signature applicability
exp/nbf/iat presence and whether exp has passed
missing recommended claims (aud, iss, jti)
sensitive data patterns in claims

A 5-part token is scanned as a JWE instead: the checks move to the encryption layer (key-management alg, content enc, CBC padding-oracle and compression risks), since there is no JWS signature to reason about.

Options

Flag	Default	Description
`-w, --wordlist <FILE>`	built-in list	Secrets to try for the weak-secret check. Falls back to a small built-in list if unset or unreadable.
`--max-crack-attempts <N>`	`100`	Cap on secrets tested during the weak-secret check.
`--skip-crack`	off	Skip the weak-secret check entirely.
`--skip-payloads`	off	Skip the attack-payload suggestions.
`--report <FILE>`	none	Write a report to a file; `.json` emits JSON, `.html`/`.htm` emits HTML. Any other extension errors.

The weak-secret check only runs on HMAC tokens (HS256/384/512). For RS/ES/PS/EdDSA it reports as not applicable. Large wordlists slow the scan down, so cap it with --max-crack-attempts during triage or CI.

Examples

Full scan, then export an HTML report for a ticket:

jwt-hack scan "$TOKEN" --report findings.html

Fast heuristics only, no cracking and no payloads:

jwt-hack scan "$TOKEN" --skip-crack --skip-payloads

CI-friendly run with a real wordlist but a bounded budget:

jwt-hack scan "$TOKEN" -w rockyou.txt --max-crack-attempts 200

Notes

Payload suggestions follow the findings: a kid header produces kid payloads, a clean token produces none. Suppress them with --skip-payloads.
The weak-secret default list is small and meant for triage. Confirm a real crack with crack and a proper wordlist.
Asymmetric algorithms flagged for confusion are a prompt to test, not a confirmed finding. Verify with payload --target alg_confusion.