Back to skill

Security audit

MGC Blackbox — Multi‑Agent Secure Execution Engine

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill benign because the flagged access, network, and local-service behaviors were judged expected for its purpose, though review confidence is medium and coverage is partial.

The verdict is benign, but it does not mean risk-free. Before installing or using the examples, consider the reported token access and local service calls, and that the weather example shares the supplied city with wttr.in. Scanning was partial: all three listed components were scanned, but several local references could not be resolved, so the evidence is incomplete.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (10)

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

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

Content

Scanner excerpt · SKILL.md (reported line 179)May include surrounding context.

python
import os, requests

# 1. Read MGC access token (fixed path, do not change)
with open(os.path.expanduser("~/.mgc/database/mgc_black_box/.mgc_token")) as f:
    MGC_TOKEN = f.read().strip()

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

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

Content

Scanner excerpt · SKILL.md (reported line 285)May include surrounding context.

mgc

text

Starts HTTP server at `<redacted-url>, initializes encrypted database on first run, generates access token at `~/.mgc/database/mgc_black_box/.mgc_token`. This token is required for all REST API calls.

## 3.4 Run MGC as an MCP server

External Transmission

Medium
Category
Data Exfiltration
Confidence
70% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 184)May include surrounding context.

md
MGC_TOKEN = f.read().strip()

# --- Usage A: Retrieve a secret value ---
resp = requests.post(
    "http://127.0.0.1:57219/api/mgc/sensitive/get",
    headers={"X-MGC-Token": MGC_TOKEN},
    json={"info_type": "token", "info_owner": "openai_api_key"},

External Transmission

Medium
Category
Data Exfiltration
Confidence
70% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 193)May include surrounding context.

md
api_key = resp.json()["data"]  # plaintext key, in memory only — never written to disk

# --- Usage B: Invoke another stored script ---
resp = requests.post(
    "http://127.0.0.1:57219/api/mgc/sensitive/get",
    headers={"X-MGC-Token": MGC_TOKEN},
    json={

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 184)May include surrounding context.

md
MGC_TOKEN = f.read().strip()

# --- Usage A: Retrieve a secret value ---
resp = requests.post(
    "http://127.0.0.1:57219/api/mgc/sensitive/get",
    headers={"X-MGC-Token": MGC_TOKEN},
    json={"info_type": "token", "info_owner": "openai_api_key"},

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 193)May include surrounding context.

md
api_key = resp.json()["data"]  # plaintext key, in memory only — never written to disk

# --- Usage B: Invoke another stored script ---
resp = requests.post(
    "http://127.0.0.1:57219/api/mgc/sensitive/get",
    headers={"X-MGC-Token": MGC_TOKEN},
    json={

Internal Network Request

Medium
Category
Server-Side Request Forgery
Confidence
70% confidence
Finding

Code issues a request to a loopback, link-local, or private-range host. This can reach internal services not meant to be exposed and is a common SSRF pivot.

Content

Scanner excerpt · SKILL.md (reported line 184)May include surrounding context.

md
MGC_TOKEN = f.read().strip()

# --- Usage A: Retrieve a secret value ---
resp = requests.post(
    "http://127.0.0.1:57219/api/mgc/sensitive/get",
    headers={"X-MGC-Token": MGC_TOKEN},
    json={"info_type": "token", "info_owner": "openai_api_key"},

Internal Network Request

Medium
Category
Server-Side Request Forgery
Confidence
70% confidence
Finding

Code issues a request to a loopback, link-local, or private-range host. This can reach internal services not meant to be exposed and is a common SSRF pivot.

Content

Scanner excerpt · SKILL.md (reported line 193)May include surrounding context.

md
api_key = resp.json()["data"]  # plaintext key, in memory only — never written to disk

# --- Usage B: Invoke another stored script ---
resp = requests.post(
    "http://127.0.0.1:57219/api/mgc/sensitive/get",
    headers={"X-MGC-Token": MGC_TOKEN},
    json={

The quick-start workflow sends the supplied city to a third-party weather service without a privacy disclosure.

Low
Category
Output Handling
Confidence
91% confidence
Finding

The example instructs users to execute a script that transmits its city argument to wttr.in. The README emphasizes encrypted execution but does not warn that this workflow sends location-related input outside the local system; encrypted script storage does not protect data transmitted by the running script.

Content

No source excerpt is available for this finding.

The trigger "Open the interface" does not specify MGC and can overlap with requests to open unrelated interfaces.

Low
Category
Trigger Abuse
Confidence
90% confidence
Finding

The trigger table maps this generic request directly to mgc_open_webui without requiring an MGC reference or an established MGC conversation context. An agent could therefore open the MGC WebUI when the user intended another application's interface.

Content

No source excerpt is available for this finding.