Back to skill

Security audit

Mingdata Dmp Auth

Security checks for vulnerabilities and agentic risk

Overview

Review recommended: this appears to be a real Mingdata DMP auth gateway, but it handles long-lived API credentials and broad authenticated DMP access with weak user-control safeguards.

Install only if you trust the publisher and need Mingdata DMP automation. Avoid pasting production AK/SK values into chat or shell commands; use the narrowest DMP credentials available, rotate any keys previously entered this way, and confirm where OpenClaw stores env.DMP_AK and env.DMP_SK before relying on the local-only security claim.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (1)

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/minri_dmp_api.py:469
Finding

Credential Exposure Through Command-Line Arguments and Console Output

Content
View full analysis
", "example": "python minri_dmp_api.py save-credentials G3N3yfBc 2kTgY33CwjPirwS07qRsrd0XbNGBPuqG" }, ensure_ascii=False, indent=2)) sys.exit(1) access_key = sys.argv[2] secret_key = sys.argv[3] credentials_path = save_credentials(access_key, secret_key) print(json.dumps({ "success": True, "message": "Credentials saved and verified successfully", "credentials_path": str(credentials_path), "access_key": access_key, "secret_key_preview": f"{secret_key[:8]}...{secret_key[-4:]}" }, ensure_ascii=False, indent=2)) ``` The credential-testing command also prints fragments of both credentials: ```python print(f" Access Key: {ak[:4]}...{ak[-4:]}") print(f" Secret Key: {sk[:4]}...{sk[-4:]}\n") ``` The Skill documentation directs the Agent to place secrets directly in command arguments: ```bash openclaw config set env.DMP_AK "your_access_key" openclaw config set env.DMP_SK "your_secret_key" openclaw gateway restart ``` It also documents the local command-line credential workflow: ```bash python scripts/minri_dmp_api.py save-credentials ``` ### Technical Analysis Command-line ...[truncated 3410 chars]
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Findings (11)

Ssd 3

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill explicitly instructs the AI to solicit AK/SK secrets via chat, then automatically store them and run configuration commands. Collecting credentials in conversational text and immediately persisting them through automated tooling significantly expands exposure risk through logs, transcripts, prompt history, misrouting, or unauthorized reuse by other skills sharing the environment.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
94% confidence
Finding

The skill describes capabilities that can read/write files, access environment variables, perform network requests, and run shell commands, yet the manifest declares no explicit tool scope or permission boundaries. This creates an over-privileged and opaque execution model where downstream agents may invoke sensitive operations without clear policy constraints or user review.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
74% confidence
Finding

The skill is designed around persistent credential storage across sessions, both via environment configuration and local files. While persistence is functionally expected for an auth gateway, it still increases blast radius because a compromised session, over-broad tool access, or another dependent skill could reuse the stored credentials without renewed user intent.

Content

Scanner excerpt · SKILL.md (reported line 99)May include surrounding context.

python minri_dmp_api.py /audience/manage/list '{"current":1,"pageSize":10}'

创建组合人群

python minri_dmp_api.py /audience/manage/combine/create '{"name":"测试人群","trackType":"MOBILE","idTypes":["MD5_IDFA"],"data":[...]}'

text

**核心功能**:

Session Persistence

Medium
Category
Rogue Agent
Confidence
74% confidence
Finding

The skill is designed around persistent credential storage across sessions, both via environment configuration and local files. While persistence is functionally expected for an auth gateway, it still increases blast radius because a compromised session, over-broad tool access, or another dependent skill could reuse the stored credentials without renewed user intent.

Content

Scanner excerpt · SKILL.md (reported line 99)May include surrounding context.

python minri_dmp_api.py /audience/manage/list '{"current":1,"pageSize":10}'

创建组合人群

python minri_dmp_api.py /audience/manage/combine/create '{"name":"测试人群","trackType":"MOBILE","idTypes":["MD5_IDFA"],"data":[...]}'

text

**核心功能**:

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The instructions say the AI will automatically execute commands to write environment variables and restart the gateway, but do not present a strong warning or explicit consent checkpoint immediately beforehand. Automatic persistence of secrets plus service restart is a sensitive state-changing action that can disrupt sessions and store credentials beyond the user's expectations.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The documentation claims credentials are stored only locally and never uploaded to the cloud, but elsewhere instructs users to save them through an OpenClaw configuration center and restart a gateway for environment injection. This inconsistent security guidance can mislead users about where secrets reside and what trust boundaries apply, increasing the risk of inappropriate secret handling.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The trigger scenarios are broad and center on common user intents like querying audiences or creating segments, without clearly constraining when this auth skill should activate or what confirmation is required. In agent environments, broad activation criteria can cause the skill to be invoked opportunistically, including prompting for or using credentials in contexts the user did not intend.

Content

No source excerpt is available for this finding.

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
80% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · scripts/minri_dmp_api.py (reported line 104)May include surrounding context.

python
print(json.dumps({
                "error": "INSECURE_CREDENTIALS_FILE",
                "message": "⚠️ 凭证文件权限不安全,其他用户可能可以读取",
                "fix": f"chmod 600 {credentials_file}"
            }, ensure_ascii=False))
            sys.exit(2)

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The CLI accepts arbitrary user-supplied endpoints and request bodies, so this 'auth/gateway' module can be used as a generic authenticated API caller for any Mingdata DMP endpoint available to the stored credentials. In a multi-skill environment, this broad capability weakens privilege separation and makes accidental or unauthorized access to sensitive DMP functions more likely if other components can invoke this script.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

Passing access keys and secret keys as positional CLI arguments can expose them through shell history, process listings, audit logs, and job runners, even if the file is later stored with safe permissions. Because this module is specifically for credential management, encouraging this usage materially increases the chance of credential disclosure.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

The natural-language documentation and user-facing descriptions in the module header are presented in a single language without offering a language/locale choice. Under the stated policy, forcing a specific language without user opt-in can be a natural-language policy violation unless clearly justified as region-specific.

Content

No source excerpt is available for this finding.