Security audit

Cheat Code

Security checks for vulnerabilities and agentic risk

Overview

The skill is purpose-aligned with an external synthesis service, but it can send broad task context to kenoodl.com while using misleading local-only privacy language and weak safeguards for returned content.

Review bucket: install only if you are comfortable with selected task context being sent to kenoodl.com. Do not use it for confidential business data, personal data, regulated data, unreleased security details, or proprietary code unless you have separately approved that disclosure. Redact more than just API keys, preview exactly what the agent will send, and treat returned synthesis as untrusted advice rather than instructions to execute.

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: instructions.md:41
Finding: Misleading Local-Only Privacy Claim Enables Disclosure of Sensitive Context
Content: View full analysis
Content-Type: application/json { "context": "" } \`\`\` ``` ### Technical Analysis The assertion that conversations, data, ideas, and context remain on the user's machine conflicts with the subsequent instruction to transmit task context to the external `kenoodl.com` service. Although the instructions prohibit sending secrets, credentials, and API keys, they do not protect other sensitive information such as: - Personal or regulated data - Proprietary business information - Confidential source-code details - Internal system architecture - Private conversation content - Unreleased product or security information The requested payload is intentionally broad and includes the problem, current thinking, previous attempts, and obstacles. An agent following these instructions may therefore send substantial conversation-derived information off-device. Credential removal alone is not an adequate data-loss-prevention control. The issue does not grant the external service additional local privileges. Its scope is the confidentiality of information included in outbound requests and any information inferable from that context. ### ...[truncated 1298 chars]
Remediation: View remediation

T09 · Insecure Skill Coding Practices

Warning

Location: instructions.md:110
Finding: Untrusted Remote Synthesis Is Integrated Without Prompt-Injection Safeguards
Content: View full analysis
Remediation: View remediation

SkillSpector

Vulnerability Patterns

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

Findings (7)

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The README uses broad promotional language such as making an agent's talents 'limitless' and telling the agent to 'deliver' without defining safe invocation boundaries, approval requirements, or disallowed use cases. In a skill that enables external HTTP access and retrieval of out-of-band information, this can encourage overuse, unsafe delegation, and unintended triggering in sensitive contexts.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The README explicitly states that the skill gives the agent access to an external service via HTTP and a runtime token, but it does not warn users that prompts, retrieved content, or other data may be transmitted off-platform. In the context of a generic skill advertised for use with 'any agent,' the absence of privacy and data-handling disclosures materially increases the risk of users sending sensitive or regulated information to a third party without informed consent.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The skill description promises effectively unlimited capabilities without defining when the skill should be invoked or what tasks it is meant to handle. In agent ecosystems, this kind of unconstrained positioning can cause over-broad routing or user prompting that funnels many requests to an external integration, increasing the chance of unnecessary data exposure and unexpected behavior.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 90% confidence
Finding: The marketing language describes the product in vague, expansive terms such as delivering breakthroughs and handling the rest, without operational constraints or safety boundaries. This can encourage indiscriminate use of the skill and obscure the fact that requests may be sent to a third-party service, which is risky in a security-sensitive agent environment.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 88% confidence
Finding: The manifest description markets the skill as making an agent 'limitless' and able to 'tell your agent what you want' without any clear scope, trigger boundaries, or behavioral constraints. In an agent ecosystem, this kind of unconstrained invocation language can encourage overbroad delegation and unsafe use of a network-enabled skill, increasing the chance that downstream instructions are treated as generally authorized.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The skill instructs the agent to send arbitrary problem context to an external service and frames it as a normal mechanism for improving reasoning. Although it says not to include secrets, the surrounding messaging strongly encourages broad context sharing and gives privacy assurances that may cause over-trust, without clearly warning that business data, proprietary information, personal data, or regulated content could still be transmitted off-device.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 91% confidence
Finding: The worked example includes a natural-language instruction in the returned content: 'Let's crack this: don't centralize the data into a static warehouse; make it live like a flock in motion,' which effectively demonstrates and directs English-only output. The file nowhere offers a language or locale choice or documents a justified locale restriction, so the skill presents a language-policy issue.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.