Back to skill

Security audit

DocsBot AI

Security checks for vulnerabilities and agentic risk

Overview

The skill mostly matches its DocsBot AI read-only purpose, but its setup instructions include an unsafe remote installer pattern that should be reviewed before installation.

Review the installer step before installing. Prefer installing the OOMOL CLI through verified vendor documentation or a package manager, and only connect DocsBot AI accounts you intend the agent to search or read through this connector.

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

Note
Location
SKILL.md:60
Finding
Unverified Remote Installer Download and Immediate Shell Execution## Vulnerability Details **File Location**: `SKILL.md`, lines 60–64 **Vulnerability Type**: Remote payload retrieval and execution **Risk Level**: Malicious ### 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 The installation instructions download mutable scripts from `cli.oomol.com` and immediately execute the returned content using Bash or PowerShell. Neither command pins an immutable release, validates a cryptographic signature or checksum, nor saves the script for inspection before execution. HTTPS protects the script while it is in transit but does not establish that its contents are safe. A compromised vendor account, hosting platform, DNS or PKI infrastructure, or malicious future replacement of the installer could cause different code to execute after the Skill has been reviewed. Installing the `oo` CLI may be necessary when it is unavailable, but piping an unverified network response directly into a shell exceeds the minimum behavior necessary to perform that installation safely. The Skill does limit this instruction to first-time setup after a command-not-found error, but that condition does not mitigate the integrity risk. ### Attack Path 1. The user or Agent attempts to use the Skill on a system where the `oo` CLI is unavailable. 2. The first-time setup instructions are followed. 3. Bash or PowerShell retrieves the installer from `cli.oomol.com`. 4. The network response is passed directly to the local shell without integrity verification or review. 5. If the hosted installer or its delivery infrastructure has been compromised, attacker-controlled commands execute immediately with the permissions of the invoking user. 6. The payload can access resources available to that user and may download further components, alter user configuration, or establish persistence. ### Impact Ass ...[truncated 894 chars]
Remediation
## Remediation Suggestions 1. Remove the `curl | bash` and `irm | iex` installation patterns. 2. Require explicit user approval before installing the CLI. 3. Prefer a trusted platform package manager with a pinned package version. 4. If a standalone installer is required: - Pin an immutable, versioned release URL. - Download the installer to a local file rather than executing the response stream. - Verify a vendor-published cryptographic signature or pinned SHA-256 digest. - Abort installation if verification fails. - Allow the user to inspect the downloaded installer before execution. 5. Execute the installer without administrative privileges unless a documented installation step strictly requires elevation. 6. Document the files, directories, environment settings, and permissions modified by installation. 7. Provide uninstall and rollback instructions. 8. Keep the existing behavior that installation occurs only after an actual command-not-found failure, rather than proactively installing software.
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
Content
- **`oo: command not found`** — install the oo CLI (other platforms: <https://cli.oomol.com/install-guide.md>):

  ```bash
  curl -fsSL https://cli.oomol.com/install.sh | bash    # macOS / Linux
  ```

  ```powershell
Confidence
98% confidence
Finding
The skill instructs the agent to install software by piping a remote script directly into a shell, which is a classic supply-chain and remote code execution risk. If the install endpoint, transport, DNS, or upstream distribution is compromised, arbitrary code would execute immediately on the host with the user's privileges.

Vague Triggers

Medium
Confidence
95% confidence
Finding
This activation guidance is broad and lacks constraints or exclusion examples, making it unclear when the skill should activate versus when a user is merely discussing DocsBot AI. Because SQP-1 applies to markdown files, this kind of ambiguous invocation wording should be narrowed to avoid unintended use.

Static analysis

No suspicious patterns detected.