Back to skill

Security audit

Dataworks Daily Monitor

Security checks for vulnerabilities and agentic risk

Overview

The skill’s monitoring purpose is coherent, but it recommends persistent Alibaba Cloud AccessKeys and automatically sends operational task details to Feishu without enough scoping controls.

Review this skill before installing. Use a RAM subaccount or short-lived credentials with read-only DataWorks permissions, avoid storing AccessKeys in shell startup files or broad app config, confirm exactly where Feishu reports are sent, and redact task names, owners, error messages, and console links if they reveal sensitive business information.

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
README.md:47
Finding

Persistent Plaintext Storage of Alibaba Cloud Credentials

Content
View full analysis

Vulnerability Details

File Location: README.md, lines 47-67
Vulnerability Type: Plaintext storage of long-lived cloud credentials
Risk Level: Medium

The documentation instructs users to export an Alibaba Cloud AccessKey ID and AccessKey secret and recommends persisting the configuration in either the OpenClaw gateway configuration file or a shell startup file.

bash
export ALIYUN_ACCESS_KEY_ID="LTAI5tXXXXXXXXXXXX"
export ALIYUN_ACCESS_KEY_SECRET="XXXXXXXXXXXXXXXXXXXX"

export DATAWORKS_PROJECT_ID="XXXXXX"
export DATAWORKS_REGION_ID="cn-shanghai"

Recommended persistent configuration locations:

text
~/.openclaw/openclaw.json
~/.bashrc

Technical Analysis

Long-lived AccessKey credentials are bearer secrets: possession is generally sufficient to authenticate as the associated Alibaba Cloud RAM identity. Persisting these credentials directly in a shell initialization file or general application configuration creates an avoidable plaintext secret at rest.

Shell startup files and gateway configuration files may be exposed to other local processes running under the same account, overly broad filesystem permissions, workstation backups, diagnostic bundles, support archives, or accidental configuration disclosure. Environment variables initialized from these files may also be inherited by child processes and exposed through process inspection or debugging facilities, depending on the operating-system configuration.

The example values are placeholders rather than actual leaked credentials. The vulnerability is the recommended credential-storage pattern.

Attack Path

  1. A user follows the documentation and writes a valid AccessKey ID and secret into ~/.bashrc or ~/.openclaw/openclaw.json.
  2. An attacker obtains local read access to the user's files, compromises another process operating as that user, or acquires a backup or diagnostic archive containing the configuration.
  3. The ...[truncated 1079 chars]
Remediation
View remediation

Remediation Suggestions

  1. Prefer short-lived Alibaba Cloud STS credentials instead of long-lived AccessKeys.
  2. Where supported, use workload identity, instance RAM roles, or another identity mechanism that avoids distributing static secrets.
  3. Store unavoidable secrets in a dedicated secret manager or operating-system credential store rather than ~/.bashrc or a general gateway configuration file.
  4. Retrieve credentials only when needed and avoid exporting them globally to every child process in the user's shell session.
  5. Assign the RAM identity only the read-only DataWorks permissions required to list and inspect task instances. Do not use an Alibaba Cloud primary-account AccessKey.
  6. If file-based storage cannot be avoided, use a dedicated credentials file outside the project, restrict its permissions to the owning user, exclude it from backups and diagnostic archives where feasible, and never commit it to version control.
  7. Add credential rotation and revocation procedures to the documentation. Rotate immediately if a configuration file or environment dump may have been disclosed.
  8. Replace the current recommendation with secret-manager or STS-based setup examples and explicitly warn against persistent plaintext storage in shell startup files.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill explicitly states it will generate reports and send them via Feishu, including failed task details and error reasons, but it does not clearly warn users that operational metadata may be transmitted to an external messaging platform. Those details can include task names, owners, error messages, and links that may reveal internal system structure or sensitive business context, creating a real data exposure risk if recipients, channels, or retention are not tightly controlled.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

For a markdown skill file, vague trigger phrases should be flagged when they are generic enough to collide with ordinary user speech. "任务运行日报" does not explicitly mention DataWorks or monitoring scope, so it could unintentionally activate on unrelated daily task-report requests.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The skill requires access to cloud credentials from environment variables but does not clearly warn users that it will consume sensitive AccessKey material to call Alibaba Cloud APIs. While reading credentials from environment variables is common, the missing disclosure and lack of least-privilege guidance increase the risk of overbroad credential exposure and unsafe deployment practices.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The skill states that it will send task reports via Feishu, but does not clearly warn the user that operational task data, failure details, and possibly sensitive job names or error messages will be transmitted to an external messaging platform. This can lead to unintended data disclosure, especially in environments where task names, error text, or links reveal internal systems or business processes.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The natural-language instructions, examples, and usage guidance are presented only in Chinese, which effectively forces a specific language for interacting with the skill. The file does not indicate that Chinese is optional, user-selectable, or required for a region-specific compliance reason.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
81% confidence
Finding

Natural-language policy violations include locale constraints imposed without user opt-in. The document sets a default region of cn-shanghai, and while this may be operationally convenient, it does not explain whether the skill is region-specific or allow the user to choose another locale/region.

Content

No source excerpt is available for this finding.