Back to skill

Security audit

Pet First Aid

Security checks for vulnerabilities and agentic risk

Overview

This pet first-aid skill is on-topic, but it needs Review because it requests unnecessary filesystem access and uses an unpinned npx installer.

Review the permission request before installing. This skill should not need filesystem access for pet first-aid advice, and the install path should ideally use a pinned, trusted ClawHub version. In real emergencies, call an emergency veterinarian or animal poison-control professional immediately and use the skill only as supplemental guidance.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • 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
Findings (2)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:15
Finding

Unpinned Package Execution Through npx

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, line 15
Vulnerability Type: T08: Insecure Dependencies
Risk Level: Medium

Vulnerable Code:

yaml
openclaw:
  requires:
    tools: [filesystem]
  install: "npx clawhub install pet-first-aid"

Technical Analysis

The installation command invokes clawhub through npx without specifying an exact package version or integrity constraint. Depending on the local npm environment, npx may retrieve the current package release from the configured package registry and execute it. The project contains no lockfile, checksum, vendored installer, or other mechanism that establishes which clawhub artifact was reviewed and trusted.

This creates a mutable supply-chain boundary: the code executed during installation may differ from the code available when the Skill was audited. Exploitation would require compromise or malicious control of the package, package publisher, registry resolution, or configured registry.

Attack Path

  1. An attacker compromises the package publisher, takes over the package, influences dependency resolution, or controls a registry configured on the victim's system.
  2. The attacker publishes or serves a malicious version of clawhub.
  3. A user follows the declared installation procedure.
  4. npx resolves and downloads the unpinned package version.
  5. Malicious package code or lifecycle behavior executes with the privileges of the user running the command.

Impact Assessment

Successful exploitation could provide arbitrary code execution under the installing user's account. The resulting access could include reading or modifying files available to that user, accessing environment variables and local credentials, making network requests, and altering the installed Skill. Administrative impact would occur only if the user executes the installation command with elevated privileges; no automatic privilege elevation is shown ...[truncated 20 chars]

Remediation
View remediation

Remediation Suggestions

  • Pin clawhub to an exact, reviewed version, for example by using an explicit version rather than an unconstrained package name.
  • Record and verify the expected package integrity hash.
  • Use a lockfile or equivalent reproducible dependency manifest where supported.
  • Configure installation to use a trusted registry and document the expected package publisher.
  • Disable or carefully review dependency lifecycle scripts where the installation workflow permits it.
  • Periodically audit the pinned release and update it through a controlled review process.

T05 · Unauthorized Access and Privilege Escalation

Note
Location
SKILL.md:14
Finding

Unnecessary Filesystem Capability Violates Least Privilege

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, line 14
Vulnerability Type: T05: Unauthorized Access and Privilege Escalation
Risk Level: Low

Vulnerable Code:

yaml
openclaw:
  requires:
    tools: [filesystem]
  install: "npx clawhub install pet-first-aid"

Technical Analysis

The Skill declares access to the filesystem tool, but its documented functionality consists of pet-emergency triage, first-aid instructions, veterinary referral information, structured conversational state, and reminders. No documented workflow requires reading from or writing to the local filesystem.

Granting an unnecessary tool broadens the authority available to the Skill and violates the principle of least privilege. The declaration is not, by itself, evidence that files are currently accessed maliciously. However, it increases the consequences of later instruction injection, unauthorized Skill modification, or runtime compromise by making filesystem operations available within the Skill's execution context.

Attack Path

  1. The Skill is installed and activated with its declared filesystem capability.
  2. An attacker introduces malicious instructions through a compromised Skill update, altered local Skill file, or another instruction-injection route affecting the runtime.
  3. The injected instructions direct the agent to invoke the filesystem tool.
  4. The agent reads or modifies files permitted by the tool's runtime scope.
  5. Retrieved information could subsequently be disclosed if another available communication channel permits it, or local files could be altered directly.

This path is conditional on a separate compromise or injection vector; the audited file does not itself contain instructions to misuse the filesystem.

Impact Assessment

The potential scope is determined by the filesystem tool's sandbox and runtime permissions. If unrestricted, exposure could include files readable or writable b ...[truncated 297 chars]

Remediation
View remediation

Remediation Suggestions

  • Remove filesystem from the required tools because the documented pet-care workflow does not need local file access.
  • If future functionality requires persistence, request filesystem access only for the specific operation that needs it.
  • Restrict any required access to a dedicated Skill data directory rather than the broader user filesystem.
  • Prefer read-only access where writes are unnecessary.
  • Enforce runtime allowlists for permitted paths and file operations.
  • Review capability declarations whenever Skill functionality changes to prevent permission creep.
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 (2)

Rp1

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The install command uses npx clawhub install pet-first-aid without pinning a specific package version, so execution depends on whatever version is currently resolved from the registry at install time. That creates a supply-chain risk: a compromised, typosquatted, or newly published breaking/malicious version of the tool could run arbitrary code during installation.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

This skill provides detailed emergency veterinary triage and treatment guidance, including poisoning, choking, seizures, and heatstroke, but does not clearly state up front that it is not a substitute for immediate professional veterinary care and may be incomplete or region-specific. In a high-stakes medical context, users may over-rely on the skill, delay urgent treatment, or apply guidance inappropriately, increasing the risk of serious harm or death to the animal.

Content

No source excerpt is available for this finding.