ClawHub
Back to skill

Security audit

Kaleido

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a read-only Kaleido connector, but its setup instructions tell users to run unverified internet installer scripts directly in a shell.

Review the installer path before using this skill. The Kaleido read actions are narrowly described, but if oo is not installed, prefer an official verifiable install method instead of running the provided curl | bash or irm | iex commands directly.

A.I.G

Vulnerability Patterns
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • 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
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Findings (1)

T03 · Remote Payload Retrieval and Execution

Error
Location
SKILL.md:66
Finding

Unverified Remote Installer Scripts Executed Directly by Shells

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 66–70
Vulnerability Type: Remote payload retrieval and execution
Risk Level: Critical

Vulnerable Code

bash
curl -fsSL https://cli.oomol.com/install.sh | bash    # macOS / Linux
powershell
irm https://cli.oomol.com/install.ps1 | iex           # Windows PowerShell

Technical Analysis

Both installation commands retrieve mutable scripts from cli.oomol.com and immediately execute the returned content with Bash or PowerShell. The instructions provide no pinned release version, checksum validation, cryptographic signature verification, or opportunity to inspect the downloaded content before execution.

HTTPS protects the connection in transit but does not establish that the script remains safe if the hosting origin, publishing account, DNS configuration, or release process is compromised. The effective executable payload can also change after this Skill has been reviewed.

Installing the oo CLI is relevant to the Skill's functionality when the command is missing. However, executing unverified network output directly is not the minimum-risk mechanism necessary to install that dependency. The Skill otherwise declares only Bash(oo *) as its allowed tool scope, while these setup instructions invoke general-purpose shells capable of performing arbitrary operations.

Attack Path

  1. A user attempts to use the Skill on a system where the oo CLI is unavailable.
  2. The resulting oo: command not found error directs the user to the first-time setup instructions.
  3. The user runs one of the documented installation commands.
  4. The command retrieves a mutable script from cli.oomol.com.
  5. Bash or PowerShell executes the response without integrity or authenticity verification beyond HTTPS.
  6. If the remote origin or distribution process has been compromised, attacker-controlled commands execute with the privileges of the user run ...[truncated 760 chars]
Remediation
View remediation

Remediation Suggestions

  1. Remove the direct curl | bash and irm | iex installation commands.
  2. Direct users to an official installation guide or package manager that supports version-pinned releases.
  3. Download the installer to a local file before execution so it can be inspected:
    • Use a fixed HTTPS URL containing an immutable release version.
    • Reject redirects to unapproved hosts.
    • Verify a publisher-provided SHA-256 or stronger checksum.
    • Prefer cryptographic signature verification against a pinned, documented signing key.
    • Execute the installer only after verification succeeds.
  4. Require explicit user approval before installing software; the agent should not automatically execute first-time setup commands after a failed connector invocation.
  5. Run installation with ordinary user privileges whenever possible and clearly document whether any elevation is required.
  6. Publish reproducible installer sources and retain immutable release artifacts so the reviewed payload corresponds to the installed payload.
  7. Preserve the existing guidance that setup should occur only after a matching failure, but replace shell execution of network output with a verifiable installation process.

SkillSpector

Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

External Script Fetching

High
Category
Supply Chain
Confidence
97% confidence
Finding

The skill recommends installing software via a remote script piped directly into a shell, which executes code fetched over the network without review or integrity verification. If the install endpoint, transport path, or hosting account is compromised, users could execute attacker-controlled code on their machine.

Content

Scanner excerpt · SKILL.md (reported line 66)May include surrounding context.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The description says to use this skill for "ANY Kaleido request," which is an extremely broad activation condition and does not define clear boundaries for when the skill should or should not be invoked. Although it is domain-scoped to Kaleido, it still lacks specificity or exclusion conditions, increasing the risk of unintended invocation for loosely related requests.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.