ClawHub
Back to skill

Security audit

glkvm

Security checks for vulnerabilities and agentic risk

Overview

The skill is purpose-aligned for GLKVM administration, but it handles full-control credentials and device actions with weakened transport and token storage safeguards.

Install only if you administer the target GLKVM device on a trusted management network. Prefer valid certificates, a pinned certificate or public key, and a private per-session cookie file; delete tokens and screenshots after use. Treat firmware, power, reset, and virtual media commands as disruptive operations requiring explicit confirmation.

A.I.G

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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)

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:19
Finding

TLS Certificate Verification Disabled for Administrative API Traffic

Content
View full analysis
" \ "https:///api/auth/login" ``` ```text - Response `ok: true` with a `token` indicates successful login; the auth_token is also saved in the cookie. - Response with `two_step_required: true` means waiting for two-step approval. - All subsequent requests must include `-b /tmp/glkvm_cookies.txt`. **All requests use HTTPS with `-k` (ignore certificate errors).** ``` ### Technical Analysis The `curl -k` option, also known as `--insecure`, disables TLS certificate-chain and hostname verification. Although traffic remains encrypted, the client does not authenticate the GLKVM endpoint. This behavior is applied to the login request and, according to the skill instructions, every subsequent authenticated API request. Consequently, any attacker capable of intercepting or redirecting network traffic can present an arbitrary certificate without causing the connection to fail. The login request transmits the fixed administrative username and the user-supplied administrator password. A successful interception can therefore expose either the administrator password or the resulting authentication token. Manipulated API responses could also mislead the agent about the state or outcome of security-sensitive operations. ### Attack Path 1. The victim invokes the skill and supplies the GLKVM IP address and administrator password. 2. An attacker obtains a network interception or redirection position, such as through a compromised gateway, malicious access point, DNS manipulation, or local network spoofing. 3. The attacker impersonates the GLKVM endpoint using an untrusted TLS certificate. 4. Because every documented request uses `-k`, certificate and hostname verification failures are ignor ...[truncated 1348 chars]
Remediation
View remediation

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
92% confidence
Finding

The command embeds a sensitive user-supplied password into a shell command and sends it via curl with insecure flags, creating tool-parameter abuse risk through unsafe handling of secrets and transport settings. Even if the IP placeholder is nominally user-provided, the broader skill context enables powerful remote control actions, so unsafe parameterization materially increases the chance of credential leakage or misuse.

Content

Scanner excerpt · SKILL.md (reported line 22)May include surrounding context.

Step 2: Login to Obtain Token

bash
curl -sk -c /tmp/glkvm_cookies.txt \
  -F "user=admin" \
  -F "passwd=<PASSWORD>" \
  "https://<IP>/api/auth/login"

External Transmission

Medium
Category
Data Exfiltration
Confidence
95% confidence
Finding

The skill instructs the agent to transmit the user's administrative password to a remote device over HTTPS while explicitly disabling certificate verification with curl -k/-s. That creates a real risk of credential interception or man-in-the-middle attacks, and the resulting token/cookie grants full control of both the GLKVM device and the attached host.

Content

Scanner excerpt · SKILL.md (reported line 22)May include surrounding context.

Step 2: Login to Obtain Token

bash
curl -sk -c /tmp/glkvm_cookies.txt \
  -F "user=admin" \
  -F "passwd=<PASSWORD>" \
  "https://<IP>/api/auth/login"

Static analysis

No suspicious patterns detected.