Security audit

Tech Tutor

Security checks for vulnerabilities and agentic risk

Overview

Tech Tutor is a coherent educational skill, but it automatically saves and reuses learning history across user and project locations in a way users should review before installing.

Install only if you are comfortable with Tech Tutor writing local learning records and consulting them later across projects. Avoid using it with confidential code or sensitive notes unless you first decide where records should be stored and are prepared to delete or manage the .tech-tutor directories yourself.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (6)

Vague Triggers

Medium

Confidence: 89% confidence
Finding: The invocation examples are broad enough to match many ordinary educational requests such as 'I want to learn about [topic]' or 'Help me understand this project,' which increases the chance of accidental or overly eager skill activation. In a skill that can search sources and create records, over-broad triggering can cause unnecessary data access, unexpected behavior, or unanticipated local side effects when users did not explicitly intend to invoke this specific skill.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The README states that learning records are created automatically and stored in user-level and project-level directories, but it does not present a prominent warning or explicit consent step before local persistence occurs. This is risky because users may unknowingly store sensitive project names, learning history, questions, or derived insights on disk, creating privacy and confidentiality exposure, especially in shared machines or sensitive repositories.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The README states that learning records are saved automatically under user-level and project-level directories, but it does not prominently warn users that conversation-derived content may be written to disk. In a tutoring/contextual learning workflow, users may paste proprietary code, internal architecture details, or personal notes, which could then persist locally without their informed consent.

Missing User Warnings

Low

Confidence: 88% confidence
Finding: The README says the skill can teach topics through real-time search of authoritative sources, implying outbound network access, but it does not clearly warn users that prompts or derived queries may be sent externally. In an educational skill, users may share confidential project context while asking for explanations, creating a risk of unintended disclosure through networked lookups.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill instructs the agent to save a learning record automatically after the session, which creates persistent on-disk storage of potentially sensitive conversation content without clear opt-in consent at the moment of collection. Even though it asks where to save the record earlier, the design treats persistence as mandatory rather than optional, increasing privacy risk and the chance of storing secrets, proprietary code details, or personal learning history unexpectedly.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The skill enables checking user-level, project-level, and cross-project learning records by default to surface reviews and connections, which can expose prior session content across contexts without a clear privacy notice or per-session consent. In a multi-project environment, this may reveal sensitive project associations, proprietary topics, or prior discussion details that the user did not expect to be consulted automatically.

Static analysis

No suspicious patterns detected.