Back to skill

Security audit

skin-pro

Security checks for vulnerabilities and agentic risk

Overview

This skill needs review because its upload code can send facial photos and authentication headers outside the advertised service without enforcing a secure destination.

Before using real credentials or photos, have the client restricted to the advertised HTTPS gateway and disable or validate redirects. Upload only photos you have permission to share, enable face-ID processing only with explicit authorization, protect saved reports, and install reviewed dependency versions. The review concern is upload containment; the artifacts do not demonstrate malicious intent.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Taint TrackingDirect Taint Flow, Variable-Mediated Taint Flow, Credential Exfiltration Chain
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Findings (14)

Tainted flow: 'headers' from os.getenv (line 76, credential/environment) → requests.post (network output)

Critical
Category
Data Flow
Confidence
93% confidence
Finding

Sending authentication headers to the intended HTTPS API is legitimate, but --base-url accepts arbitrary destinations and plaintext HTTP without validation. If an attacker can influence that argument, the request exposes the access key, generated signature, and submitted facial data to an attacker-controlled destination; automatic redirects can also forward these custom authentication headers across origins. The secret key itself is not transmitted directly.

Content

Scanner excerpt · scripts/call_skin_pro.py (reported line 80)May include surrounding context.

python
"AccessKey": access_key,
        "Signature": generate_signature(access_key, secret_key),
    }
    resp = requests.post(
        url, headers=headers, data=data, files=files, timeout=timeout
    )
    resp.raise_for_status()

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The primary API call and image-processing workflow match the declared purpose. However, --face-id-detect causes the code to submit faceIdDetect=true and a required userGroup to the API, exposing an identity-related detection capability beyond the described skin analysis. The supplied code does not establish whether the server stores identities or performs identity matching. Credential loading, response filtering, and optional JSON output are supporting implementation details rather than mismatches.

Content

No source excerpt is available for this finding.

Tainted flow: 'files' from open (line 120, file read) → requests.post (network output)

High
Category
Data Flow
Confidence
80% confidence
Finding

File contents flow to a network sink. This may indicate data exfiltration of sensitive files.

Content

Scanner excerpt · scripts/call_skin_pro.py (reported line 80)May include surrounding context.

python
"AccessKey": access_key,
        "Signature": generate_signature(access_key, secret_key),
    }
    resp = requests.post(
        url, headers=headers, data=data, files=files, timeout=timeout
    )
    resp.raise_for_status()

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · scripts/call_skin_pro.py (reported line 165)May include surrounding context.

python
try:  # pragma: no cover - optional dependency
        from dotenv import load_dotenv

        load_dotenv(os.path.join(os.path.dirname(__file__), ".env"))
        load_dotenv()  # also pick up CWD .env
    except ImportError:
        pass

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · scripts/call_skin_pro.py (reported line 166)May include surrounding context.

python
from dotenv import load_dotenv

        load_dotenv(os.path.join(os.path.dirname(__file__), ".env"))
        load_dotenv()  # also pick up CWD .env
    except ImportError:
        pass

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · scripts/call_skin_pro.py (reported line 175)May include surrounding context.

python
if not ak or not sk:
        raise SystemExit(
            "Missing credentials. Set env vars SKIN_PRO_AK / SKIN_PRO_SK "
            "(or define them in a .env file). They must NOT be passed as CLI args."
        )
    return ak, sk

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · scripts/requirements.txt (reported line 3)May include surrounding context.

text
requests>=2.28
pycryptodome>=3.18
python-dotenv>=1.0  # optional: auto-load .env for local dev

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: env, file_write, network.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

The skill exposes optional face-identity detection beyond its stated skin-analysis purpose.

Medium
Category
Excessive Agency
Confidence
92% confidence
Finding

When --face-id-detect is supplied, the code sends faceIdDetect=true and a userGroup to the API; these options are exposed at L195-L196. Requesting identity-related facial processing is distinct from assessing skin condition, and the manifest does not declare this capability.

Content

No source excerpt is available for this finding.

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
96% confidence
Finding

The open-ended requests constraint does not establish a reproducible, reviewed dependency version and allows older vulnerable releases as well as unreviewed future releases. Because this skill sends facial images to an API, HTTP-client integrity matters, although the manifest alone does not establish an exploitable installation.

Content

Scanner excerpt · scripts/requirements.txt (reported line 1)May include surrounding context.

text
requests>=2.28
pycryptodome>=3.18
python-dotenv>=1.0  # optional: auto-load .env for local dev

Unverifiable Dependency: requests has 16 known advisory(ies) (CVE-2014-1830 (Exposure of Sensitive Information to an Unauthorized Actor in Requests); CVE-2024-47081 (Requests vulnerable to .netrc credentials leak via malicious URLs); CVE-2024-35195 (Requests `Session` object does not verify requests after making first request wi) +13 more), but the manifest does not pin a version, so it is unknown whether the installed release is affected

Low
Category
Supply Chain
Confidence
95% confidence
Finding

The constraint permits requests releases affected by the cited TLS-verification and .netrc credential-disclosure vulnerabilities, although it does not prove that an affected release is installed. If vulnerable versions and triggering request patterns are present, credentials or sensitive facial-image traffic could be exposed; this overlaps with the unpinned-dependency finding rather than establishing a separate confirmed exploit.

Content

No source excerpt is available for this finding.

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
96% confidence
Finding

The pycryptodome lower bound leaves the installed cryptographic implementation uncontrolled and permits releases with known vulnerabilities. This is a dependency-management weakness, but its practical impact depends on the installed version and which cryptographic operations the implementation actually uses.

Content

Scanner excerpt · scripts/requirements.txt (reported line 2)May include surrounding context.

text
requests>=2.28
pycryptodome>=3.18
python-dotenv>=1.0  # optional: auto-load .env for local dev

Unverifiable Dependency: pycryptodome has 4 known advisory(ies) (CVE-2018-15560 (PyCryptodome integer overflow vulnerability); CVE-2023-52323 (PyCryptodome and pycryptodomex side-channel leakage for OAEP decryption); CVE-2018-15560 (PyCryptodome before 3.6.6 has an integer overflow in the data_len variable in AE) +1 more), but the manifest does not pin a version, so it is unknown whether the installed release is affected

Low
Category
Supply Chain
Confidence
92% confidence
Finding

The minimum permitted pycryptodome version, 3.18, precedes the 3.19.0 fix for CVE-2023-52323, so the constraint admits an affected release. The older integer-overflow advisory is excluded by this minimum, and exploitation of the remaining issue depends on relevant OAEP-decryption usage and attacker observation capabilities, neither of which is demonstrated here.

Content

No source excerpt is available for this finding.

Unverifiable Dependency: python-dotenv has 2 known advisory(ies) (CVE-2026-28684 (python-dotenv: Symlink following in set_key allows arbitrary file overwrite via ); CVE-2026-28684 (python-dotenv reads key-value pairs from a .env file and can set them as environ)), but the manifest does not pin a version, so it is unknown whether the installed release is affected

Low
Category
Supply Chain
Confidence
40% confidence
Finding

OSV returned advisories for this package, but its resolved version is unknown. A package-name lookup does not establish whether the installed release is affected by those advisories.

Content

No source excerpt is available for this finding.