Security audit

Bytesagain Comic Script

Security checks for vulnerabilities and agentic risk

Overview

The skill is not clearly malicious, but it needs review because its comic-writing name and manifest do not match the content-marketing commands it documents and implements.

Review this package before installing. Confirm whether you want a comic scripting helper or a content-marketing CLI, and be aware that topics you pass to the documented commands may be saved locally in a history log.

A.I.G

Vulnerability Patterns

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
Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
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 (8)

Tp4

High

Category: MCP Tool Poisoning
Confidence: 99% confidence
Finding: The skill metadata advertises a comic storyboard assistant, but the documented commands implement a broader content-marketing/blog-writing tool and also log command history to local storage. This mismatch can mislead users and higher-level agents about the skill’s actual behavior and data handling, which is dangerous because trust and routing decisions may be made on the declared purpose rather than the real capabilities.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

High

Category: Not specified by scanner
Confidence: 99% confidence
Finding: The manifest says this is a comic storyboard/dialogue skill, but the body describes a generic content marketing assistant with drafting, SEO, CTA, scheduling, and repurposing functions. Such semantic deception increases the chance that users or orchestrating agents invoke the skill under false assumptions, potentially exposing prompts, topics, or workflow data to unexpected logging or misuse.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

High

Category: Not specified by scanner
Confidence: 99% confidence
Finding: The manifest says this skill drafts comic storyboards with panels, dialogue, and scene pacing, but the implemented commands are for generic content creation and optimization such as headlines, SEO, CTAs, metrics, and content schedules. This is a semantic mismatch between the declared comic-writing purpose and the actual functionality exposed by the script.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: Branding the skill as 'Comic Script' while documenting generic content/SEO workflows creates a deceptive interface, even if no direct code execution or exfiltration is shown in this file. In agent ecosystems, misleading titles and headers can cause inappropriate tool selection and overtrust, which makes other hidden or undocumented behaviors more likely to be missed.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The file comment labels the tool as a 'Content creation and optimization assistant', which conflicts with the manifest's specific purpose of drafting comic storyboards and comic-writing structure. This is an active documentation-versus-intent divergence rather than a mere omission.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The user-facing help text advertises generic content commands like headline generation, SEO tips, repurposing, and metrics, which contradict the manifest's stated comic-storyboarding role. Because this documentation actively presents a different intent to users, it constitutes intent-code divergence with the declared skill purpose.
Content: No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: Capabilities such as SEO advice, calls to action, metrics tracking, and content scheduling are not obviously required to design comic panels, write dialogue, or structure comic chapters. These features extend into marketing/content-optimization functions outside the skill's stated purpose.
Content: No source excerpt is available for this finding.

Missing User Warnings

Low

Category: Not specified by scanner
Confidence: 93% confidence
Finding: User-supplied input is written to a persistent local history log without clear disclosure or consent. In a content-creation context, prompts may contain unpublished ideas, customer data, internal campaign details, or other sensitive text, so silent retention increases privacy and data exposure risk on shared systems or through later log access.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.