Back to skill

Security audit

Openclaw Api Reference

Security checks for vulnerabilities and agentic risk

Overview

This is a plain OpenClaw API reference, but it documents deletion, skill execution, configuration changes, and unpinned SDK installs without safety boundaries or confirmation guidance.

Review this skill before installing it. It appears to be documentation, not an executable payload, but agents using it may learn how to delete OpenClaw data, create or run skills, change configuration, and install SDK packages. Use it only with least-privilege API keys, require confirmation before DELETE or skill-run actions, and pin or verify SDK package versions before installation.

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

T08 · Insecure Dependencies

Warning
Location
SKILL.md:264
Finding

Unpinned Third-Party SDK Installation Instructions

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 264 and 282
Vulnerability Type: Unpinned dependencies from public package registries
Risk Level: Medium

Vulnerable Code

bash
npm install openclaw-sdk
bash
pip install openclaw

Technical Analysis

The documentation instructs users to install third-party packages from public npm and Python package registries without specifying reviewed versions, lockfiles, integrity hashes, or trusted source links. Consequently, the installed package is determined by mutable registry state at installation time rather than by the audited Skill content.

Package installation can execute lifecycle hooks, including npm installation scripts. Installed packages may also execute code when subsequently imported by the documented JavaScript or Python examples. If either package or its publisher account is compromised, ownership changes, or an unsafe release is published, following these instructions could introduce unreviewed code into the user's environment.

The available evidence does not establish that the named packages are currently malicious. The risk arises from unconstrained dependency resolution and insufficient supply-chain verification.

Attack Path

  1. An attacker compromises the publisher account or release process for openclaw-sdk or openclaw, or otherwise causes a malicious release to become the version selected by the package manager.
  2. The attacker publishes a release containing a malicious installation hook or runtime payload.
  3. A user follows the instructions in SKILL.md and runs the unpinned npm install or pip install command.
  4. The package manager retrieves the attacker-controlled release from the public registry.
  5. Malicious code executes during installation where applicable, or when the package is imported through the documented SDK example.
  6. The payload operates with the privileges of the user or automation account performing the installation or runn ...[truncated 639 chars]
Remediation
View remediation

Remediation Suggestions

  1. Pin each dependency to an exact version that has been reviewed, for example:
    bash
    npm install --save-exact openclaw-sdk@REVIEWED_VERSION
    pip install openclaw==REVIEWED_VERSION
    
  2. Publish and maintain lockfiles where these dependencies are used, such as package-lock.json or a hash-locked Python requirements file.
  3. For Python, require artifact hashes with pip install --require-hashes.
  4. For npm, retain lockfile integrity metadata and use npm ci in automated environments.
  5. Document the packages' official registry pages and source repositories so users can verify publisher identity and provenance.
  6. Review package contents, transitive dependencies, and npm lifecycle scripts before recommending a release.
  7. Use automated dependency scanning and monitor publisher ownership, release provenance, and newly disclosed vulnerabilities.
  8. Perform installation in a least-privileged, isolated environment without production credentials or unnecessary filesystem access.
Vulnerability Patterns
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • 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
Findings (4)

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · SKILL.md (reported line 114)May include surrounding context.

删除会话

bash
DELETE /api/v1/sessions/:id

消息 API

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · SKILL.md (reported line 133)May include surrounding context.

删除消息

bash
DELETE /api/v1/messages/:id

Skills API

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The document exposes destructive endpoints for deleting sessions and messages without any caution about confirmation, authorization scope, backups, or irreversible effects. In a skill/reference context, this can normalize unsafe use and increase the chance that an agent or developer invokes deletion operations without appropriate safeguards.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill documents creating and executing skills, including a script field with shell content and parameterized execution, but provides no warning that these capabilities may lead to arbitrary code execution or unsafe tool invocation. In agent-facing documentation, this is especially risky because consumers may treat the examples as approved workflows and pass untrusted scripts or parameters directly into execution.

Content

No source excerpt is available for this finding.