Back to skill

Security audit

TDengine Database

Security checks for vulnerabilities and agentic risk

Overview

The skill is a small TDengine query helper, but it documents a hardcoded root database login for read-only market-data queries.

Review before installing. Replace the documented root login with a dedicated read-only TDengine account, do not pass passwords on the command line, and install Python dependencies in an isolated environment with pinned versions or a trusted package source.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (2)

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:7
Finding

Hardcoded Privileged TDengine Credentials

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, line 7
Vulnerability Type: Hardcoded plaintext credentials and excessive database privileges
Risk Level: High

Vulnerable Code Snippet:

bash
taos -h localhost -u root -p taosdata

Technical Analysis

The documented connection command exposes both the privileged TDengine username (root) and its plaintext password (taosdata). Anyone able to read the Skill can recover and attempt to reuse these credentials. Supplying the password as a command-line argument can also expose it through shell history, process listings, logs, or command auditing systems.

Use of the database root account violates least-privilege principles because the documented operations are read-only market-data queries and do not require unrestricted administrative access.

Attack Path

  1. An attacker or untrusted Agent reads SKILL.md.
  2. The attacker extracts the disclosed username and password.
  3. If the documented credentials remain active, the attacker runs taos -h localhost -u root -p taosdata from an environment able to reach the service.
  4. The attacker authenticates with root-level database privileges.
  5. The attacker enumerates accessible databases and performs any operations permitted to the TDengine root account, potentially including reading, modifying, or deleting data and changing database configuration or users.

Exploitation depends on the credentials remaining valid and the attacker having network or local access to the TDengine service.

Impact Assessment

Successful exploitation may grant administrative TDengine access rather than only the read access needed by the documented queries. The affected scope can include all databases, tables, users, and administrative operations available to the root account. Potential consequences include confidential market-data disclosure, unauthorized modification, destructive deletion, service disruption, and creati ...[truncated 38 chars]

Remediation
View remediation

Remediation Suggestions

  • Immediately rotate the exposed password and verify that the documented value is not active in any environment.
  • Create a dedicated read-only TDengine account restricted to the exact databases, tables, and query operations required by this Skill.
  • Store credentials in an approved secret manager or protected environment variable rather than in project documentation.
  • Avoid passing passwords directly on the command line. Use a secure client configuration, secret-file mechanism with restrictive permissions, or interactive password prompt.
  • Review shell histories, CI logs, process telemetry, and documentation copies for prior credential exposure.
  • Restrict TDengine network access to approved hosts and monitor privileged login attempts.

T08 · Insecure Dependencies

Warning
Location
SKILL.md:10
Finding

Unpinned Third-Party Package Installation

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, line 10
Vulnerability Type: Unverified and non-reproducible dependency installation
Risk Level: Medium

Vulnerable Code Snippet:

bash
pip install taospy

Technical Analysis

The installation instruction specifies neither an audited package version nor cryptographic hashes. Consequently, the package artifact installed by this command can change over time according to the configured Python package indexes and dependency resolution state.

The command also does not identify an approved package source. A compromised package release, compromised configured index, or unsafe index precedence could therefore introduce unreviewed code into the execution environment. Python packages and their dependencies may execute code during installation or when imported.

This finding does not establish that taospy is currently malicious; it identifies an unsafe dependency acquisition practice that lacks version and artifact integrity controls.

Attack Path

  1. A user or Agent follows the instruction in SKILL.md.
  2. pip resolves taospy and its transitive dependencies from the environment's configured package indexes.
  3. If a resolved release or index has been compromised, pip downloads an attacker-controlled or otherwise unreviewed artifact.
  4. Malicious code executes during package installation, initialization, or subsequent import.
  5. The code runs with the permissions of the user or automation account that invoked pip.

Exploitation requires compromise or manipulation of a resolved package artifact, dependency, or configured package source.

Impact Assessment

A malicious dependency would execute with the invoking process's privileges. Depending on those privileges and environment access, it could read local files and secrets, alter application data, tamper with the Python environment, make network connections, or establish further compromise. Ins ...[truncated 90 chars]

Remediation
View remediation

Remediation Suggestions

  • Pin taospy and every transitive dependency to reviewed versions in a lock file.
  • Record and verify cryptographic hashes using a hash-locked requirements file and pip install --require-hashes.
  • Download packages only from an explicitly approved and authenticated internal mirror or trusted index.
  • Review package provenance, maintainers, release history, and artifacts before approving updates.
  • Install dependencies inside an isolated virtual environment or container under a non-privileged account.
  • Add automated dependency vulnerability and integrity scanning to the update process.
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
Findings (1)

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The markdown includes field descriptions such as '超级表', '具体表', and 'TAG字段' entirely in Chinese, with no indication that the skill is intentionally limited to Chinese-speaking users or a China-specific deployment context. Under the stated policy, forcing a specific language without user opt-in or justification is a natural-language policy issue.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.