Security audit

Esfj Coach

Security checks for vulnerabilities and agentic risk

Overview

This coaching skill appears to mix ESFJ packaging with INTP coaching content and asks the agent to keep ongoing personal growth records without clear user consent or retention controls.

Review before installing. Only use this skill if you are comfortable with potentially mismatched personality coaching and with the agent saving personal growth summaries. Ask the publisher to align the ESFJ/INTP materials and make memory strictly opt-in with clear review, edit, delete, and retention controls.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (16)

Intent-Code Divergence

High

Confidence: 99% confidence
Finding: The README describes an INTP-focused coaching skill while the manifest declares an ESFJ coaching skill, creating a clear identity and behavior mismatch. This can cause the wrong prompt logic, safety framing, and user expectations to be applied, which is especially risky in a guidance-oriented skill where users may rely on personality-specific advice.

Description-Behavior Mismatch

High

Confidence: 98% confidence
Finding: The documented behavior is explicitly tailored to INTP traits and problems rather than the ESFJ-focused purpose declared in the metadata. In a coaching skill, this mismatch can mislead users into receiving inappropriate psychological framing and recommendations, undermining trust and potentially causing harmful or irrelevant advice.

Description-Behavior Mismatch

High

Confidence: 97% confidence
Finding: The manifest claims this is an ESFJ coaching skill, but the description, keywords, and tags describe an INTP-focused coaching skill. This inconsistency can mislead users, reviewers, and downstream routing systems, causing the wrong skill to be invoked or trusted under false pretenses; in a coaching context, that can produce inappropriate guidance and undermine provenance checks.

Intent-Code Divergence

High

Confidence: 99% confidence
Finding: The reference content labeled as ESFJ describes traits and function dynamics that align with a different MBTI profile, creating a core instruction/data integrity failure. In a coaching skill, this can systematically generate misleading guidance for vulnerable users seeking personalized advice, undermining trust and potentially reinforcing the wrong behavioral interventions.

Intent-Code Divergence

Medium

Confidence: 95% confidence
Finding: The file metadata points to a different skill path (`intp-coach/...`) than the declared ESFJ skill, indicating cross-skill contamination or copy-paste drift. This is dangerous because it increases the likelihood that the runtime or maintainers will load, trust, or preserve incorrect background knowledge, compounding the misalignment in user-facing coaching.

Description-Behavior Mismatch

High

Confidence: 98% confidence
Finding: The skill manifest says this skill is for ESFJ growth coaching, but the referenced material is entirely about INTP cognition, pitfalls, and coaching strategy. This mismatch can cause the agent to deliver systematically incorrect personality-specific guidance, undermining user trust and potentially causing harmful advice in a mental-wellbeing or self-development context because recommendations are tailored to the wrong profile.

Vague Triggers

Medium

Confidence: 85% confidence
Finding: Using 'automatic detection' for mode selection without clear trigger boundaries or exclusion conditions makes the skill's behavior unpredictable. In a coaching context, ambiguous routing can cause users seeking reflective dialogue to receive directive advice, or vice versa, which can degrade safety, reliability, and informed user control.

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: The README says the skill saves conversations and action records automatically, but it does not disclose retention details or obtain user consent. Because coaching conversations may contain sensitive emotional or mental-health-adjacent information, silent persistence creates privacy and compliance risk if users do not understand what is stored and why.

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The skill explicitly promises a 'growth record' and future recall of prior conversations, which implies retention of personal coaching data without any clear privacy notice, consent flow, retention limit, or opt-out. In a coaching context, users are likely to share sensitive emotional, behavioral, and career information, so silent persistence increases privacy and trust risks.

Missing User Warnings

Medium

Confidence: 97% confidence
Finding: This section instructs ongoing logging of user problems, interests, actions, and 'aha moments' but does not require informing the user that these details will be stored. Because these records can reveal psychological patterns and personal goals, undisclosed collection materially increases privacy exposure.

Missing User Warnings

Medium

Confidence: 98% confidence
Finding: The conversation-closing workflow mandates archiving key user information automatically, again without user warning or permission. Auto-archiving at the end of a supportive coaching conversation is especially risky because users may reasonably assume the exchange is ephemeral unless told otherwise.

Ssd 3

Medium

Confidence: 96% confidence
Finding: Persistent collection of conversation details into a user profile creates a durable behavioral record that can be misused, over-retained, or exposed in a breach. The skill is framed as personal growth coaching, making the stored data potentially sensitive even if it is not explicitly medical.

Ssd 3

Medium

Confidence: 95% confidence
Finding: The workflow requires storing key user insights during coaching conversations for future use, which is persistent profiling. Since coaching insights can include vulnerabilities, motivations, failures, and decision patterns, this persistence meaningfully raises privacy and profiling risk.

Ssd 3

Medium

Confidence: 98% confidence
Finding: The dedicated growth-record section formalizes ongoing logging of personal problems, interests, actions, and insights, amounting to sustained user profiling. In aggregate, this can reveal intimate patterns about the user's mental state, habits, and life direction, increasing harm if accessed improperly.

Ssd 3

Medium

Confidence: 97% confidence
Finding: A mandatory rule to record user profile information for future conversations normalizes persistent storage as a required behavior rather than a user choice. That is dangerous because it bakes privacy-invasive profiling into the skill's core operation without safeguards or transparency.

Ssd 3

Medium

Confidence: 98% confidence
Finding: The end-of-conversation flow directs archiving of user problems, insights, action plans, and later execution results, creating a longitudinal dossier of the user's growth journey. Such persistent accumulation is especially sensitive in a coaching setting and can magnify harm through unauthorized access, secondary use, or unexpected recall.

Static analysis

No suspicious patterns detected.