GHSA-X92V-RPX6-P6CW

Vulnerability from github – Published: 2026-06-18 13:58 – Updated: 2026-06-18 13:58
VLAI
Summary
PraisonAI: Webhook signature verification skipped (fail-open) when secret unset, allowing forged inbound webhooks (WhatsApp & Linear bots)
Details

The WhatsApp and Linear bot adapters verify the inbound webhook HMAC signature only when a secret is configured. When the secret environment variable is unset — the default on a fresh install and common in development — verification is skipped entirely and the webhook body is parsed and dispatched as a genuine, trusted event. A remote, unauthenticated attacker who can reach the bot's webhook endpoint can inject arbitrary platform events.

Affected code:

WhatsApp - src/praisonai/praisonai/bots/whatsapp.py - init (line 108): self._app_secret = app_secret or os.environ.get("WHATSAPP_APP_SECRET", "") -> defaults to "" - route (line 246): app.router.add_post(self._webhook_path, self._handle_webhook) -> default path "/webhook" - _handle_webhook (lines 585-595): if self._app_secret: gates the ENTIRE check; when falsy the body is json.loads()'d and dispatched to _process_webhook_data() with no verification.

Linear - src/praisonai/praisonai/bots/linear.py - init (line 86): self._signing_secret = signing_secret or os.environ.get("LINEAR_WEBHOOK_SECRET", "") -> "" - _handle_webhook (lines 244-248): same if self._signing_secret: fail-open guard. - start() (lines 169-170): only logs a warning; does not fail closed.

The _verify_signature implementations themselves are correct (constant-time HMAC-SHA256); the defect is that verification is bypassed when the secret is absent.

Impact: - WhatsApp: attacker POSTs a crafted Meta Cloud API payload spoofing any sender and message text; injected into agent sessions and processed as a real user message (prompt injection, unauthorized agent/command invocation, contact impersonation). - Linear: attacker POSTs forged AgentSession / Comment events, causing the agent to act on and comment on issues no legitimate event referenced. The webhook routes require no other authentication, so exploitation needs only network reachability.

Proof of concept (bot started without the secret - the default):

curl -X POST http://VICTIM:PORT/webhook \ -H 'Content-Type: application/json' \ -d '{"object":"whatsapp_business_account","entry":[{"changes":[{"value": {"messages":[{"from":"15551234567","id":"wamid.x","type":"text", "text":{"body":"attacker-injected message"}}]}}]}]}' # No X-Hub-Signature-256 header; bot returns 200 and processes the message. # Linear: omit LINEAR_WEBHOOK_SECRET and POST without a Linear-Signature header.

A self-contained PoC that executes the real _handle_webhook / _verify_signature source extracted from the repo confirms: secret unset -> status 200, payload dispatched (VULNERABLE); secret set + no signature -> status 403, nothing dispatched (control).

Remediation: Fail closed. When no secret is configured, reject all webhooks (HTTP 403) and refuse to start the adapter unless a secret is set (or an explicit, clearly-named insecure-dev override is given):

if not self._app_secret: return web.Response(status=403, text="Webhook secret not configured") signature = request.headers.get("X-Hub-Signature-256", "") if not self._verify_signature(body, signature): return web.Response(status=403, text="Invalid signature")

Distinct from prior advisories: The accepted default-insecure advisories cover a different surface/mechanism — CALL_SERVER_TOKEN unset (GHSA-86qc-r5v2-v6x6) and the JWT key default "dev-secret-change-me" (GHSA-3qg8-5g3r-79v5). This is in the bot webhook adapters and the mechanism is skipping signature verification entirely when the secret is absent, not a weak default key.

Show details on source website

{
  "affected": [
    {
      "database_specific": {
        "last_known_affected_version_range": "\u003c= 4.6.52"
      },
      "package": {
        "ecosystem": "PyPI",
        "name": "praisonai"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "4.6.59"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [],
  "database_specific": {
    "cwe_ids": [
      "CWE-345",
      "CWE-347"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-06-18T13:58:08Z",
    "nvd_published_at": null,
    "severity": "HIGH"
  },
  "details": "The WhatsApp and Linear bot adapters verify the inbound webhook HMAC signature only\nwhen a secret is configured. When the secret environment variable is unset \u2014 the\ndefault on a fresh install and common in development \u2014 verification is skipped entirely\nand the webhook body is parsed and dispatched as a genuine, trusted event. A remote,\nunauthenticated attacker who can reach the bot\u0027s webhook endpoint can inject arbitrary\nplatform events.\n\nAffected code:\n\nWhatsApp - src/praisonai/praisonai/bots/whatsapp.py\n- __init__ (line 108): self._app_secret = app_secret or os.environ.get(\"WHATSAPP_APP_SECRET\", \"\")  -\u003e defaults to \"\"\n- route (line 246): app.router.add_post(self._webhook_path, self._handle_webhook)  -\u003e default path \"/webhook\"\n- _handle_webhook (lines 585-595): `if self._app_secret:` gates the ENTIRE check; when falsy the body is\n  json.loads()\u0027d and dispatched to _process_webhook_data() with no verification.\n\nLinear - src/praisonai/praisonai/bots/linear.py\n- __init__ (line 86): self._signing_secret = signing_secret or os.environ.get(\"LINEAR_WEBHOOK_SECRET\", \"\")  -\u003e \"\"\n- _handle_webhook (lines 244-248): same `if self._signing_secret:` fail-open guard.\n- start() (lines 169-170): only logs a warning; does not fail closed.\n\nThe _verify_signature implementations themselves are correct (constant-time HMAC-SHA256);\nthe defect is that verification is bypassed when the secret is absent.\n\nImpact:\n- WhatsApp: attacker POSTs a crafted Meta Cloud API payload spoofing any sender and message\n  text; injected into agent sessions and processed as a real user message (prompt injection,\n  unauthorized agent/command invocation, contact impersonation).\n- Linear: attacker POSTs forged AgentSession / Comment events, causing the agent to act on and\n  comment on issues no legitimate event referenced.\nThe webhook routes require no other authentication, so exploitation needs only network\nreachability.\n\nProof of concept (bot started without the secret - the default):\n\n  curl -X POST http://VICTIM:PORT/webhook \\\n    -H \u0027Content-Type: application/json\u0027 \\\n    -d \u0027{\"object\":\"whatsapp_business_account\",\"entry\":[{\"changes\":[{\"value\":\n         {\"messages\":[{\"from\":\"15551234567\",\"id\":\"wamid.x\",\"type\":\"text\",\n         \"text\":{\"body\":\"attacker-injected message\"}}]}}]}]}\u0027\n  # No X-Hub-Signature-256 header; bot returns 200 and processes the message.\n  # Linear: omit LINEAR_WEBHOOK_SECRET and POST without a Linear-Signature header.\n\nA self-contained PoC that executes the real _handle_webhook / _verify_signature source\nextracted from the repo confirms: secret unset -\u003e status 200, payload dispatched (VULNERABLE);\nsecret set + no signature -\u003e status 403, nothing dispatched (control).\n\nRemediation:\nFail closed. When no secret is configured, reject all webhooks (HTTP 403) and refuse to start\nthe adapter unless a secret is set (or an explicit, clearly-named insecure-dev override is given):\n\n  if not self._app_secret:\n      return web.Response(status=403, text=\"Webhook secret not configured\")\n  signature = request.headers.get(\"X-Hub-Signature-256\", \"\")\n  if not self._verify_signature(body, signature):\n      return web.Response(status=403, text=\"Invalid signature\")\n\nDistinct from prior advisories:\nThe accepted default-insecure advisories cover a different surface/mechanism \u2014 CALL_SERVER_TOKEN\nunset (GHSA-86qc-r5v2-v6x6) and the JWT key default \"dev-secret-change-me\" (GHSA-3qg8-5g3r-79v5).\nThis is in the bot webhook adapters and the mechanism is skipping signature verification entirely\nwhen the secret is absent, not a weak default key.",
  "id": "GHSA-x92v-rpx6-p6cw",
  "modified": "2026-06-18T13:58:08Z",
  "published": "2026-06-18T13:58:08Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/MervinPraison/PraisonAI/security/advisories/GHSA-x92v-rpx6-p6cw"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/MervinPraison/PraisonAI"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:H/A:L",
      "type": "CVSS_V3"
    }
  ],
  "summary": "PraisonAI: Webhook signature verification skipped (fail-open) when secret unset, allowing forged inbound webhooks (WhatsApp \u0026 Linear bots)"
}



Log in or create an account to share your comment.




Tags
Taxonomy of the tags.


Loading…

Loading…

Loading…

Forecast uses a logistic model when the trend is rising, or an exponential decay model when the trend is falling. Fitted via linearized least squares.

Sightings

Author Source Type Date Other

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or observed by the user.
  • Confirmed: The vulnerability has been validated from an analyst's perspective.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
  • Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
  • Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
  • Not confirmed: The user expressed doubt about the validity of the vulnerability.
  • Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.

Loading…

Detection rules are retrieved from Rulezet.

Loading…

Loading…