Security audit

Psyke SEO & GEO Audit

Security checks for vulnerabilities and agentic risk

Overview

The skill is coherent for a branded SEO/GEO audit, but it directs public publishing of client audit material without a clear consent or review gate.

Install only if you are comfortable with a Psyke-branded workflow that may crawl websites, use external AI/search tools, create a public GitHub repository, and deploy the audit publicly to Vercel. Before use, require explicit client approval, review staged files, prefer private/access-controlled publishing, and pin deployment tooling.

A.I.G

Vulnerability Patterns

Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Insecure DependenciesIntroduces malicious components through unsafe dependency sources
Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
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

Findings (3)

T01 · Skill Instruction Hijacking

Error

Location: references/slide-structure.md:61
Finding: Mandatory Promotional Content and Lead-Generation Links in Audit Output
Content: View full analysis
Remediation: View remediation

T08 · Insecure Dependencies

Warning

Location: SKILL.md:185
Finding: Unpinned Remote Package Execution Through npx
Content: View full analysis
Remediation: View remediation

T09 · Insecure Skill Coding Practices

Warning

Location: SKILL.md:184
Finding: Forced Public Repository Can Disclose Client Audit Data
Content: View full analysis
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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 (6)

Missing User Warnings

High

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The skill directs public GitHub publication and Vercel deployment without an explicit warning or consent step, which is especially dangerous because the generated artifact may include client names, URLs, audit findings, and competitive analysis. In this business context, that can leak sensitive commercial information and reputationally damaging security or SEO weaknesses.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The trigger phrases are broad enough to match routine requests like 'run an audit' or 'audit this site', increasing the chance that the skill activates unintentionally. Because the skill performs external fetching, benchmarking, and publication steps, accidental invocation can lead to unapproved data collection and disclosure.
Content: No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The skill instructs creating a public GitHub repository for each audit, which can expose client audit content, site findings, competitor data, and potentially sensitive business intelligence to the public internet. This exceeds the core task of generating an audit deck and creates an unnecessary data disclosure path.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 70% confidence
Finding: npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.
Content: No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: Deploying the generated audit to Vercel publishes client-facing content to an external service, potentially making confidential findings accessible via a public URL. In the context of SEO/GEO audits, the deck may contain weaknesses, benchmarks, and internal recommendations that should not be exposed without informed consent.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: The instructions require including queries with "Australia" and Australia-specific location examples, which imposes a locale constraint in the workflow. The document does not offer a user choice of region or explain that the framework is intended only for Australian clients, so this is a natural-language locale policy issue.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.