Back to skill

Security audit

Entj Coach

Security checks for vulnerabilities and agentic risk

Overview

This coaching skill contains no executable payload, but it needs Review because its package identity is inconsistent and it directs automatic retention of personal coaching records without clear user controls.

Review carefully before installing. Do not follow the README install command unless the package identity is corrected to the reviewed skill, and only use the memory/profile feature if the host provides explicit opt-in, review, deletion, and retention controls for saved coaching data.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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)

T08 · Insecure Dependencies

Warning
Location
README.md:22
Finding

Package Identity Mismatch Can Redirect Installation to an Unreviewed Skill

Content
View full analysis
Remediation
View remediation

other

Warning
Location
README.md:50
Finding

Automatic Retention of Sensitive Coaching Profiles Lacks Consent and Lifecycle Controls

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (21)

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The declared ENTJ scope conflicts with documented behavior, examples, and guidance centered on INTP users. This is dangerous because the system may invoke the skill under false assumptions and provide personality-specific advice to the wrong users, which is especially risky in coaching contexts where users may rely on tailored guidance to make personal, academic, or career decisions.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The README describes an INTP-focused coaching skill while the manifest metadata declares an ENTJ-specific skill. This scope mismatch can cause the agent to trigger in the wrong contexts, provide guidance optimized for the wrong user population, and undermine operator trust in what the skill actually does. In a coaching or mental-wellness adjacent context, misrouting and persona confusion can materially degrade advice quality.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The reference file is explicitly written for INTP coaching while the skill metadata says the skill should only trigger for ENTJ users. This mismatch can cause the agent to deliver guidance tailored to the wrong personality profile, undermining system integrity and potentially causing harmful or misleading advice in growth, career, or learning contexts.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The inline documentation repeatedly identifies the material as 'INTP Coach' reference content, directly contradicting the declared ENTJ-only purpose of the skill. This increases the likelihood that downstream prompts or agent behavior will use incompatible reasoning patterns and advice templates, leading to incorrect personalization and erosion of trust.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The README states that conversation and action records are saved automatically, but it does not clearly explain retention, purpose, storage location, access controls, or user consent. In a coaching setting, these records can contain sensitive personal, emotional, or career-related information, so undisclosed persistence creates meaningful privacy and compliance risk.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The manifest says the skill should not trigger for users who identify as non-ENTJ, but later guidance explicitly allows service to other types. This mismatch can lead to unanticipated activation and advice outside the skill’s intended specialization, undermining informed user expectations and safe routing.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The manifest says the skill should not trigger for users who identify as non-ENTJ, but later guidance explicitly allows service to other types. This mismatch can lead to unanticipated activation and advice outside the skill’s intended specialization, undermining informed user expectations and safe routing.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill’s body expands behavior beyond the manifest’s stated trigger boundaries by engaging vague emotional inputs and non-ENTJ users. This creates scope drift: the agent may activate in situations users and platform policy would not expect, increasing the chance of unintended coaching, misrouting, and collection of sensitive personal context.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill advertises a 'growth record' that saves dialogue and action history, but this data handling is not disclosed in the manifest description. Undisclosed memory/persistence is risky because users may reveal sensitive mental-health, career, or behavioral information without realizing it may be retained for future sessions.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill describes storing user conversation details and action records without a clear user-facing privacy notice or consent mechanism. Because the content concerns growth, emotions, habits, and career decisions, retained data may be sensitive and create privacy and trust risks if stored unexpectedly.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The instructions mandate recording user profile information and archiving it across sessions without warning the user about retention. Persistent profiling of behavioral patterns and personal goals can expose sensitive inferences over time, especially in a coaching context where users may disclose vulnerabilities.

Content

No source excerpt is available for this finding.

Ssd 3

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill instructs persistent collection and reuse of user conversation details in a profile for future sessions. This is a substantive stateful data-retention behavior that can accumulate sensitive personal history and increase harm if mishandled, leaked, or used without the user's informed consent.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The end-of-conversation flow requires archiving user information automatically, but does not tell the user that their data will be stored. Silent archival at session close is particularly problematic because users may assume the interaction ends without persistence, while the system retains plans, insights, and outcomes.

Content

No source excerpt is available for this finding.

Ssd 3

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill repeatedly directs the agent to archive conversation content, action plans, and execution results for later recall. Repeated archival reinforces long-term profiling and creates a larger sensitive dataset about the user’s emotions, goals, and behavior than is necessary for a single coaching interaction.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The provided skill context says this skill should trigger for ENTJ users seeking growth/career/learning guidance. However, the manifest keywords and tags emphasize INTP/"intp" on L07-L09, which semantically describes a different audience and behavior scope than the stated ENTJ-specific purpose.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

This markdown file is entirely written in Chinese and is described as background knowledge for the entj-coach skill, but it does not indicate that Chinese is optional, user-selected, or limited to a justified locale-specific deployment. Under the policy rule, a skill that effectively constrains language without opt-in should be flagged.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The reference material for an ENTJ coaching skill appears to describe a different cognitive-function stack and mismatched traits, which can cause the agent to deliver guidance based on the wrong personality model. In a coaching context, this is dangerous because users may receive systematically incorrect self-development or career advice under false pretenses, reducing trust and potentially worsening decisions.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

This markdown file is entirely Chinese-language content, but it does not state that the skill is Chinese-only, region-specific, or offer any user language choice. Under the policy rule, forcing a specific language without user opt-in is a natural-language policy concern.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
74% confidence
Finding

The title and description position the skill specifically as an 'INTP Coach' and frame its guidance around one personality-type label. This creates a natural-language targeting constraint analogous to a fixed user-category policy without offering users a choice of framework or clarifying that the guidance is optional and not restricted to one labeled group.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Low
Category
Not specified by scanner
Confidence
72% confidence
Finding

The manifest identifies the skill as "entj-coach," but the repository URL points to a project named "women-awakening-museum," which does not match the declared skill purpose. This suggests the manifest may misrepresent the actual underlying skill package or content source.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
99% confidence
Finding

The metadata states a document path under a different skill identity, which indicates copy-paste or packaging inconsistency. While not directly exploitable as code execution, it increases the risk of the wrong reference set being loaded, maintained, or trusted, which can propagate incorrect behavior in the skill.

Content

No source excerpt is available for this finding.