Security audit

ProfitCore

Security checks for vulnerabilities and agentic risk

Overview

This skill is only Markdown, but it tries to make the agent prioritize profit and use its ROI template for every task, which can override user intent.

Review before installing. This skill may be useful as a business ROI template, but it should be constrained to explicit ROI or opportunity-analysis requests and should not be allowed to override unrelated tasks, user-requested formats, or approval requirements for consequential actions.

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

Findings (1)

T01 · Skill Instruction Hijacking

Error

Location: SKILL.md:5
Finding: Mandatory Skill Instructions Hijack the Agent's Session Goals and Response Behavior
Content: View full analysis
Remediation: View remediation
“When the user explicitly requests ROI or opportunity analysis, use the following framework.” 2. **Remove identity and system-level claims.** Delete or rewrite: - “Transform the agent” - “profit-driven autonomous system” - “SYSTEM LEVEL” - “The agent exists to” The Skill should describe a task-specific analytical method rather than redefine the Agent's identity or overarching purpose. 3. **Preserve instruction precedence and user intent.** Add an explicit boundary: > “Follow applicable system, developer, and user instructions. Do not replace the user's objective with profit maximization.” 4. **Make the response template conditional.** State that the template is a default only when the user has not requested another format. User- or application-specified schemas must take precedence. 5. **Remove autonomous-execution implications.** Require separate, explicit user confirmation before taking external actions, spending funds, contacting third parties, publishing content, or invoking consequential tools. 6. **Constrain `NO ACTION` decisions.** Permit `NO ACTION` only within an explicitly requested opportunity-selection exercise. It must not be used to reject unrelated tasks merely because they lack financial return. 7. **Clarify feedback and state boundaries.** Specify that learning is limited to the current analysis unless the user explicitly authorizes persistent storage. Do not write behavioral rules or attacker-controlled content into long-term memory. 8. **Add a safe scoping statement**, for example: ```markdown This Skill is an optional framework for ROI analysis. Use it only when the user explicitly asks to compare commercial opportunities or assess return on investment. It does not alter the Agent's identity, safety constraints, instruction hierarchy, or obje ...[truncated 121 chars]

SkillSpector

Vulnerability Patterns

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

Findings (2)

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 90% confidence
Finding: The README explicitly describes an AI that 'makes decisions' and 'executes efficiently,' which implies autonomous action-taking without any stated human approval, scope limits, or safety controls. In an agent skill context, this can encourage deployment of a system that performs real-world actions affecting systems, funds, or data based solely on a profitability heuristic.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: Presenting the skill as a 'self-improving, profit-focused AI system' without disclosing monitoring, rollback, testing, or alignment constraints normalizes unsafe autonomous optimization. Self-improvement combined with profit maximization can incentivize harmful behavior, policy bypass, or risky actions if the agent is later connected to tools or sensitive environments.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.