Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB
CVE-2026-59967 Overview
CVE-2026-59967 is an Insecure Direct Object Reference (IDOR) vulnerability in Tandoor Recipes, an open-source application for managing recipes, meal planning, and shopping lists. The flaw affects versions prior to 2.6.10 and resides in the POST /api/recipe-from-source/ endpoint. The endpoint looks up BookmarkletImport records by sequential primary key within the active Space without validating that the requesting user is the record owner. An authenticated member of a shared Space can submit another member's pending bookmarklet identifier to retrieve the victim's captured URL and page content, then trigger an unconditional deletion of that pending import. The issue is tracked under [CWE-639: Authorization Bypass Through User-Controlled Key] and is fixed in version 2.6.10.
Critical Impact
Authenticated Space members can read and destroy other members' pending bookmarklet imports, exposing private captured page content and causing data loss.
Affected Products
- Tandoor Recipes versions prior to 2.6.10
- Shared Space deployments with multiple authenticated members
- Instances exposing the /api/recipe-from-source/ endpoint
Discovery Timeline
- 2026-10-16 - CVE-2026-59967 published to NVD
- 2026-10-19 - Last updated in NVD database
Technical Details for CVE-2026-59967
Vulnerability Analysis
Tandoor Recipes stores pending bookmarklet captures in a BookmarkletImport database table keyed by a sequential primary key. When a user submits a bookmarklet to POST /api/recipe-from-source/, the server resolves the record using only the primary key and the active Space. The server returns the stored url and captured html fields in the response, then deletes the record.
Because the lookup does not restrict results to created_by=request.user, any authenticated member of the same Space can supply another member's bookmarklet ID. The server treats the attacker as the legitimate owner, exposes the victim's captured page content, and removes the record from the database.
Root Cause
The root cause is a missing ownership check in the queryset filter. The vulnerable code filtered BookmarkletImport objects only by primary key, relying on Space membership for authorization rather than per-user ownership. This matches the [CWE-639] pattern of authorization decisions made on predictable, user-controlled identifiers.
Attack Vector
Exploitation requires an authenticated account that shares a Space with the victim. Primary keys are sequential integers, making enumeration straightforward. The attacker iterates candidate IDs through the recipe-from-source endpoint and parses the response for the victim's captured URL and HTML. Each successful request also destroys the victim's pending import, producing a secondary integrity and availability impact.
if serializer.is_valid():
if (b_pk := serializer.validated_data.get('bookmarklet', None)) and (
- bookmarklet := BookmarkletImport.objects.filter(pk=b_pk).first()):
+ bookmarklet := BookmarkletImport.objects.filter(pk=b_pk, space=request.space, created_by=request.user).first()):
serializer.validated_data['url'] = bookmarklet.url
serializer.validated_data['data'] = bookmarklet.html
bookmarklet.delete()
Source: GitHub Commit 18e13a3. The patch adds space=request.space and created_by=request.user to the queryset filter, enforcing per-user ownership.
Detection Methods for CVE-2026-59967
Indicators of Compromise
- Repeated POST /api/recipe-from-source/ requests from a single user containing sequential or enumerated bookmarklet identifier values.
- Unexpected disappearance of pending BookmarkletImport records belonging to other Space members.
- Audit log entries showing bookmarklet imports deleted shortly after creation by an account other than the creator.
Detection Strategies
- Monitor application access logs for high-volume requests to /api/recipe-from-source/ with varying bookmarklet parameter values from a single session.
- Correlate BookmarkletImport creation and deletion events, flagging cases where the deleting user differs from created_by.
- Add database-layer auditing on the BookmarkletImport table to track read and delete operations against records not owned by the acting user.
Monitoring Recommendations
- Enable verbose request logging on the Tandoor Recipes API gateway or reverse proxy, including request body parameters.
- Alert on any Space member performing deletions on records owned by other users within shared Spaces.
- Review historical traffic for enumeration patterns against the vulnerable endpoint prior to patching.
How to Mitigate CVE-2026-59967
Immediate Actions Required
- Upgrade Tandoor Recipes to version 2.6.10 or later, which enforces created_by and space filtering on BookmarkletImport lookups.
- Audit shared Spaces for untrusted members and revoke access where ownership of multi-tenant data cannot be verified.
- Review recent BookmarkletImport activity for signs of enumeration or unauthorized deletions.
Patch Information
The fix is included in Tandoor Recipes 2.6.10. The relevant code change is documented in commit 18e13a3 and in GitHub Security Advisory GHSA-4vw7-c646-g23w. The patch adds space=request.space, created_by=request.user to the BookmarkletImport.objects.filter() call in cookbook/views/api.py.
Workarounds
- Restrict Space membership to trusted users until the upgrade is applied.
- Place the Tandoor Recipes instance behind an authenticated reverse proxy and rate-limit requests to /api/recipe-from-source/ to slow enumeration.
- Instruct users to avoid saving sensitive URLs or page content through the bookmarklet feature until patched.
# Upgrade Tandoor Recipes via Docker Compose
docker compose pull
docker compose up -d
# Verify the running version is 2.6.10 or later
docker exec tandoor-recipes python -c "import cookbook; print(cookbook.__version__)"
Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB

