Security audit

Comic Script

Security checks for vulnerabilities and agentic risk

Overview

The package is advertised as a comic storyboard helper, but its main documented CLI is a content-marketing tool that also keeps a local command history.

Review this skill carefully before installing. It does not consistently match its comic-script description, and one included CLI behaves like a content-marketing assistant. Avoid entering sensitive drafts or proprietary ideas unless you are comfortable with command arguments being stored in a local 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 (9)

Tp4

High

Category: MCP Tool Poisoning
Confidence: 99% confidence
Finding: The skill’s manifest and title claim a comic storyboard assistant, but the documented behavior is a generic terminal content-marketing/blog-writing tool with SEO, CTA, scheduling, and repurposing features. This mismatch is dangerous because users and higher-level agents may grant trust, route tasks, or apply permissions based on the declared purpose while the skill performs unrelated actions, including local activity logging, creating a transparency and trust-boundary failure.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

High

Category: Not specified by scanner
Confidence: 99% confidence
Finding: The command list documents capabilities for drafting blog content, headlines, SEO, schedules, hooks, CTAs, and metrics rather than comic panels, dialogue, pacing, or chapter structure. In an agent ecosystem, this semantic deception can cause incorrect tool selection and unsafe trust decisions, especially when the declared identity implies a narrow creative use case but the actual scope is broader and includes persistent logging.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

High

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The usage guidance and examples repeatedly steer users toward general content creation workflows like blog drafting, newsletters, SEO, and repurposing, directly contradicting the stated comic-storyboarding context. This increases the risk of misuse and misclassification by users or orchestration systems, though the impact is somewhat lower here because this section reinforces the mismatch rather than introducing new capabilities beyond the already documented scope.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

High

Category: Not specified by scanner
Confidence: 99% confidence
Finding: The inline description explicitly labels the script as a content creation and optimization assistant, directly contradicting the manifest's comic-script positioning. Such contradictory identity metadata is a strong indicator of repurposed or mislabeled code, which undermines user trust and can conceal unauthorized functionality behind an innocuous skill name.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

High

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The help text documents a generic content-optimization tool, not a comic storyboard skill, so the user-facing interface itself misrepresents what the skill does. In a skill ecosystem, misleading help output can cause incorrect trust decisions, accidental disclosure of content prompts, and misuse of a tool under a false category.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

High

Category: Not specified by scanner
Confidence: 99% confidence
Finding: The implemented commands and help text describe a generic content-marketing assistant rather than a comic storyboard/script tool. This mismatch is dangerous because users and orchestrators may grant the skill trust, routing, or data based on the manifest, while the actual behavior performs unrelated functions, enabling deceptive capability substitution and misuse.
Content: No source excerpt is available for this finding.

Context-Inappropriate Capability

High

Category: Not specified by scanner
Confidence: 98% confidence
Finding: SEO, CTA, metrics, scheduling, and repurposing features are unrelated to a comic storyboard drafting skill and materially expand the tool's behavior beyond its declared scope. This is risky because unjustified capabilities increase the chance of deceptive packaging, inappropriate invocation, and collection or processing of user content for purposes the user did not intend.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: User-supplied input is appended to a local history log without disclosure, which creates an unannounced persistence channel for potentially sensitive prompts, story ideas, or proprietary content. In this skill context, users may enter unpublished creative material, making silent logging more dangerous because it can expose confidential drafts to other local users, backups, or later compromise.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 92% confidence
Finding: The draft command emits a generic word-count template instead of comic-oriented output such as panels, scenes, and dialogue. In context, this reinforces that the skill is not performing its advertised function, which can mislead downstream users or systems into invoking an unfit tool under false pretenses.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.