Back to skill

Security audit

Alibaba Cloud Database RDS Supabase

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed Alibaba Cloud RDS Supabase management guide, with expected cloud-credential and API use, but users should keep permissions narrow and confirm sensitive changes.

Install only if you intend to let an agent help manage Alibaba Cloud RDS Supabase resources. Use a least-privilege RAM role or short-lived credentials, avoid exposing the full credentials file in prompts or logs, redact passwords and secrets from saved outputs, and require explicit confirmation before destructive or security-sensitive operations.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • 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)

T05 · Unauthorized Access and Privilege Escalation

Warning
Location
SKILL.md:25
Finding

Broad Access to Long-Lived Alibaba Cloud Credentials

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 25-29
Vulnerability Type: Excessive access to a user-level credential store
Risk Level: Medium

Vulnerable Code Snippet:

markdown
## AccessKey Priority (Required)

1) Environment variables (preferred):`ALICLOUD_ACCESS_KEY_ID` / `ALICLOUD_ACCESS_KEY_SECRET` / `ALICLOUD_REGION_ID`
Region policy: `ALICLOUD_REGION_ID` is optional default; if unset choose the most reasonable region and ask when unclear.  
2) Standard credentials file:`~/.alibabacloud/credentials`

Technical Analysis

The Skill explicitly directs the Agent to use ~/.alibabacloud/credentials as a credential source. This user-level file may contain long-lived AccessKey secrets and multiple profiles with permissions extending beyond the requested RDS Supabase task.

Authentication to Alibaba Cloud is necessary for the declared functionality, and the Skill appropriately prefers environment variables and recommends least-privilege RAM credentials. However, instructing an Agent to access the credential file directly expands the sensitive-data boundary. If the file is read as ordinary text rather than through an official SDK credential provider, its complete contents could enter the Agent context, command output, logs, or generated evidence.

No instruction to exfiltrate, print, or persist credentials was found, and the documented external destinations are official Alibaba Cloud domains. The risk is therefore excessive credential exposure and potential misuse rather than confirmed credential theft.

Attack Path

  1. A user invokes the Skill for an Alibaba Cloud RDS Supabase operation.
  2. Preferred environment variables are unavailable.
  3. Following the required credential priority, the Agent accesses ~/.alibabacloud/credentials.
  4. Direct file inspection exposes one or more long-lived AccessKey pairs to the Agent execution context.
  5. Malicious instructions, accidental logging, ...[truncated 1003 chars]
Remediation
View remediation

Remediation Suggestions

  1. Remove instructions that encourage direct reading or printing of ~/.alibabacloud/credentials.
  2. Require the official Alibaba Cloud SDK credential-provider chain so secrets are resolved without being placed in prompts, command output, evidence files, or logs.
  3. Prefer short-lived STS credentials or an assumed RAM role over long-lived AccessKeys.
  4. Scope RAM permissions to the required RDS AI operations, resource identifiers, and regions. Separate read-only and mutating roles where practical.
  5. Require explicit user confirmation before destructive or security-sensitive operations, including deletion, password resets, public endpoint changes, IP allowlist changes, and SSL disablement.
  6. Explicitly prohibit saving AccessKey secrets, database passwords, dashboard passwords, storage secrets, private keys, or unredacted sensitive API responses under output/aliyun-rds-supabase/.
  7. Redact secret-bearing fields from logs and evidence, including AccessKeySecret, password fields, storage credentials, certificate private keys, authorization headers, signatures, and session tokens.
  8. If file-based profiles must remain supported, accept only a user-selected profile through the SDK provider and enforce restrictive filesystem permissions without reading the entire file into the Agent context.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (6)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill explicitly advertises destructive and security-sensitive capabilities such as deleting instances, resetting passwords, and modifying IP whitelist/SSL/auth settings, but it does not require explicit user confirmation before mutating actions. In an agent setting, this increases the chance of accidental or socially engineered destructive changes to production resources.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The document is written entirely in Chinese and begins immediately with Chinese-language headings and instructions, with no indication that users may choose another language or that the skill is region-specific. Under the policy, forcing a specific language without user opt-in is a natural-language policy concern.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% 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 · references/sources.md (reported line 4)May include surrounding context.

md
官方文档来源(用于后续更新)
============================

- https://api.aliyun.com/document/RdsAi/2025-05-07/overview
- https://www.alibabacloud.com/help/zh/rds/apsaradb-rds-for-postgresql/api-rdsai-2025-05-07-overview-postgresql
- https://www.alibabacloud.com/help/zh/rds/apsaradb-rds-for-postgresql/api-rdsai-2025-05-07-dir-postgresql/
- https://www.alibabacloud.com/help/zh/rds/apsaradb-rds-for-postgresql/supabase/

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The skill instructs use of access keys via environment variables or credential files but does not warn against echoing, logging, persisting, or including secrets in evidence artifacts. Because the skill also directs saving outputs and command results to disk, there is a realistic risk that credentials or sensitive API data could be exposed in logs or output files.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
94% confidence
Finding

SQP-3 applies to all file types and includes language or locale policy violations. This markdown file presents all user-facing content exclusively in Chinese, with no indication that the user can choose another language or that the content is intentionally limited to a Chinese-speaking audience.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

The heading is written entirely in Chinese ('官方文档来源(用于后续更新)'), and the file provides no indication that language selection is optional or that the skill is intentionally restricted to a Chinese-speaking or China-region context. Under the policy, fixed language/locale usage without opt-in or documented justification is a natural-language policy concern.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.