ClawHub
Back to skill

Security audit

ChessMaster

Security checks for vulnerabilities and agentic risk

Overview

This chess skill is transparent about its game API use, but it requires persistent token storage and unattended background polling that users should review before installing.

Install only if you are comfortable with an agent storing game bearer tokens and polling the external chess service in the background. Prefer using a secure credential store, deleting tokens when a game ends, avoiding logs or ordinary memory for tokens, and requiring explicit user approval before unattended heartbeat play.

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 (2)

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:165
Finding

Bearer Token Persisted Without Mandatory Security Controls

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
HEARTBEAT.md:46
Finding

Heartbeat Instructions Permit Bearer Token Storage in a Generic File or Database

Content
View full analysis
Remediation
View remediation

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (7)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The skill explicitly instructs the agent to persistently store agentToken values and to automatically use them for authenticated network calls, but it provides no safeguards around secret handling, storage protection, rotation, minimization, or user consent. In an agent ecosystem, long-lived tokens in persistent memory increase the chance of credential leakage, replay, or cross-task misuse if memory is exposed or another skill can access it.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
83% confidence
Finding

The file includes explicit instructions to transmit bearer tokens to an external service, which is expected for the chess integration but still constitutes external data transmission of sensitive credentials. While the destination appears aligned with the skill's purpose, the danger comes from normalizing outbound authenticated requests without any trust boundary explanation, domain validation guidance, or warning that tokens are being sent off-platform.

Content

Scanner excerpt · HEARTBEAT.md (reported line 25)May include surrounding context.

1. Check Game Status

bash
curl https://chessmaster.mrbean.dev/api/agents/game/YOUR_ROOM_ID \
  -H "Authorization: Bearer YOUR_AGENT_TOKEN"

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
88% confidence
Finding

The instruction to operate without external human intervention authorizes autonomous decision-making and external actions after the initial trigger. In context this is game play rather than financial or system administration activity, so the danger is reduced, but it still creates a capability for unsupervised network actions and combines with the heartbeat/persistence instructions to extend activity beyond a single user request.

Content

Scanner excerpt · SKILL.md (reported line 159)May include surrounding context.

md
## Operational Guidelines

1. **Autonomously Execute Moves**: Once a game starts, operate without external human intervention.
2. **Handle Time Constraints**: Respond within the `timeLimit`. Game loss occurs if `lives` are exhausted.
3. **Stale Room Timeout**: If no move is made for **30 minutes**, the game will automatically end as a draw due to inactivity. Keep your heartbeat running!
4. **Respect Invitations**: Access games only via provided `roomId`; room listing is not supported.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill explicitly instructs the agent to store the agentToken and roomId in persistent memory or a database so it can reconnect later. Persisting reusable authentication material beyond the active session increases the chance of credential leakage, unintended reuse across contexts, or later access without fresh user approval; this is broader than what is necessary for a simple chess interface.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The documentation tells the agent to persist the authentication token but provides no warning about credential sensitivity, retention, or privacy implications. That omission normalizes unsafe credential handling and increases the risk that tokens are stored insecurely or retained longer than necessary.

Content

No source excerpt is available for this finding.

Ssd 3

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The natural-language instruction to save authentication tokens for later reuse creates a direct credential retention risk. If the agent's memory, logs, or backing database are exposed, the token could let an attacker re-enter games, submit moves, or monitor game state without authorization from the user at the time of access.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The heartbeat section directs autonomous periodic execution and checking of active games without a user prompt. This expands the skill from a user-invoked chess interface into ongoing background behavior, which can bypass user awareness, trigger unrequested external actions, and create a standing channel for continued interaction with the remote service.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.