Security audit

Rune Wisdom

Security checks across malware telemetry and agentic risk

Overview

This is a static rune divination skill, but it includes under-qualified medical, coercive, and harmful occult instructions that warrant Review before installation.

Install only if you specifically want an esoteric rune-divination reference and are comfortable with the model seeing and potentially using practical occult, medical-adjacent, coercive, and curse-related material. Treat all health content as non-medical folklore, do not use it for diagnosis or treatment, and avoid using the skill for identifying or accusing real people, crisis situations, consent-related questions, or harmful magical requests.

SkillSpector

By NVIDIA

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection

Findings (38)

Context-Inappropriate Capability

Medium

Confidence: 95% confidence
Finding: The file gives concrete quasi-medical treatment instructions for specific conditions such as fever, infections, skin disease, dizziness, and immune issues, including timing, materials, and contraindications. Even though framed as historical or esoteric reference material, these passages can be acted on as health advice and may delay proper medical care or encourage ineffective treatment.

Context-Inappropriate Capability

Medium

Confidence: 93% confidence
Finding: The file includes formulas and interpretations for influencing another person's emotions or consent, including statements that a desired person will not refuse the practitioner. This normalizes coercive manipulation and can be used to support emotional abuse, stalking, or non-consensual influence under a spiritual pretext.

Context-Inappropriate Capability

Medium

Confidence: 88% confidence
Finding: The file documents behavior such as fetching external websites and synthesizing new rune entries, which expands the skill from a static reference role into network-backed acquisition and autonomous content generation. In an agent context, that broadens the trust boundary and can enable prompt injection from remote sources, unreviewed data incorporation, and drift from curated reference material.

Context-Inappropriate Capability

Low

Confidence: 80% confidence
Finding: The inclusion of absolute local filesystem paths reveals implementation details about the environment and project layout that are unrelated to the user-facing rune functionality. While not directly exploitable on its own, such path disclosure can aid reconnaissance, leak usernames or directory structure, and encourage an agent to reason about or attempt access to local files outside its intended scope.

Context-Inappropriate Capability

High

Confidence: 98% confidence
Finding: The protocol instructs the agent to provide health prognoses, grave outcome assessments, and triage-like warnings based on divination, including possible fatal outcomes and paralysis. That crosses from spiritual entertainment into pseudo-medical guidance, which can delay real care, intensify distress, or cause users to act on false health claims.

Context-Inappropriate Capability

High

Confidence: 97% confidence
Finding: The skill claims it can detect curses, magical attacks, harmful actors, and similar hidden threats through rune draws. This can reinforce paranoia, delusions, coercive beliefs, or interpersonal conflict by presenting unverifiable supernatural accusations as actionable insight.

Context-Inappropriate Capability

High

Confidence: 99% confidence
Finding: The protocol directs the agent to identify real-world people from rune results and assign harmful roles or traits to them, such as being destructive, deceptive, or responsible for wrongdoing. This is dangerous because it can produce false accusations, reputational harm, stalking or retaliation, and manipulation of the user's personal relationships.

Intent-Code Divergence

Medium

Confidence: 89% confidence
Finding: The file contains conflicting safety logic: it says to refuse readings in active crisis, but elsewhere instructs the agent to deliver grave prognoses and explicit danger warnings in those same domains. This inconsistency makes unsafe behavior more likely because the model receives mixed operational guidance about whether to disengage and refer out or continue escalating divinatory claims.

Vague Triggers

Medium

Confidence: 89% confidence
Finding: The skill’s trigger scope is broad enough to activate on vague requests about runes, spirituality, or Norse/Germanic topics, which can cause the agent to provide divination or occult guidance when the user may have intended historical, linguistic, or cultural information instead. In this context, overbroad routing increases the chance of unsolicited spiritual, magical, or quasi-advisory content being delivered without clear user consent or framing.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The file provides divination, magical, and healing-oriented instructions, including talismans, rune healing, and practical applications, without a prominent disclaimer that this is spiritual or entertainment content and not medical, legal, or factual advice. In practice, users could rely on this material for health, life, or decision-making guidance, especially because the text presents procedures and recommendations in an authoritative tone.

Missing User Warnings

Medium

Confidence: 97% confidence
Finding: The document presents ritual healing claims in an instructional format without a clear, repeated warning that they are occult beliefs rather than evidence-based medicine. In a user-facing skill, this omission increases the chance that users treat the guidance as safe substitute care for real medical issues.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The file provides extensive ritual, breathing, chanting, posture, and meditation instructions intended for practical use, while only offering a narrow warning about one rune possibly invoking a demonic entity. This selective warning can mislead users into assuming the broader practice set is safe, despite risks such as psychological distress, coercive magical framing, and misuse for health or interpersonal manipulation.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: This content presents operational magical effects such as necromancy, battle-protection, and influencing another person's affection as applicable practices without any safety, mental-health, fraud, or ethical disclaimer. Even if framed as religious or esoteric material, packaging these as actionable capabilities can encourage delusional beliefs, manipulative conduct, or risky real-world decisions in vulnerable users.

Missing User Warnings

High

Confidence: 99% confidence
Finding: The table explicitly normalizes coercive relationship manipulation by listing effects like 'win a maiden's affection' and 'secure reluctant maiden' without any consent warning. This is dangerous because it frames non-consensual influence as a valid outcome, which can reinforce stalking, harassment, or emotional abuse behaviors under a spiritual or magical justification.

Missing User Warnings

High

Confidence: 99% confidence
Finding: This file provides extensive disease-specific treatment protocols, contraindications, dosing/timing instructions, diagnostic guidance, and therapeutic claims for real medical conditions without a prominent warning that the content is historical/esoteric and not safe medical advice. A user could reasonably treat it as actionable healthcare guidance, delay appropriate care, or follow unsafe recommendations such as treating inflammatory, endocrine, cardiovascular, reproductive, or renal conditions with rune rituals, herbal infusions, or dietary restrictions.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: This section presents coercive and harmful ritual practices in a practical, instructional framing without any safety warning, ethical boundary, or clear limitation to literary/historical analysis. Even if the claims are occult rather than scientifically grounded, the content normalizes threats, compulsion, and adversarial targeting in a way that can encourage harmful behavior or coercive intent.

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: This portion enumerates actionable harmful techniques such as curse construction, banishment, poisoning-adjacent enchanted drinks, sexual binding, and methods for incapacitating or targeting others, again without safety boundaries. In the context of a skill explicitly positioned for runic magic and divination, these descriptions can be repurposed as pseudo-instructions for coercive abuse, stalking, intimidation, or ritualized self-justification for harming others.

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The document explicitly states that the article contains manipulative and harmful magical applications and frames them as intentional operative practice. Even though there is some high-level cautionary language, it is not a clear safety boundary and does not prevent the content from being used as guidance for coercive or destructive acts against others.

Missing User Warnings

High

Confidence: 99% confidence
Finding: This section contains concrete harmful-use instructions, including limiting a target's control over material life and advising use in gambling or high-risk all-or-nothing scenarios. These are actionable directives rather than neutral scholarship, making the skill more likely to facilitate manipulation, fraud-adjacent behavior, or harmful occult coercion narratives.

Missing User Warnings

High

Confidence: 100% confidence
Finding: Across this span, the file repeatedly provides operational suggestions for cursing, suppressing will, inducing dependency, destroying relationships, undermining judgment, and harming a target. In the context of a rune divination and magic skill, this is more dangerous because the content is in-scope for likely user requests and can be surfaced as practical advice, not merely archival reference.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: This section gives operational instructions for occult practices aimed at influencing other people, applying symbols to consumables or bodies, and destroying carriers afterward, but provides no safety framing, consent boundaries, or disclaimer that these are spiritual beliefs rather than reliable methods. In a user-facing agent, that can normalize coercive behavior, unsafe substitution for real-world help, or harmful disposal practices, especially because the skill presents the instructions as practical procedure rather than folklore or fiction.

Missing User Warnings

Medium

Confidence: 97% confidence
Finding: The file instructs the reader to give explicit warnings of fatal illness, catastrophic injury, or family rupture based on rune combinations, and even says not to soften the prognosis. In an interactive skill about divination, this is dangerous because users may treat the output as genuine medical or life-safety guidance, causing panic, delayed care, reckless decisions, or psychological harm.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The file gives practical guidance on using human or animal blood in rune work without any health, biohazard, self-harm, or legal safety warning. Even in an esoteric reference, normalizing blood use can encourage unsafe handling, self-injury, or exposure to bloodborne pathogens.

Missing User Warnings

Medium

Confidence: 97% confidence
Finding: This section provides actionable magical formulas framed as bindings, compulsion, deception, curses, disruption, and harm against other people. Regardless of the metaphysical claims, the content operationalizes coercive and abusive intent and can be used to justify stalking, manipulation, intimidation, or retaliatory behavior.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The file repeatedly gives actionable ritual instructions involving open flames and, elsewhere in the document, lying near candles or placing objects on the body, but it lacks a prominent global safety disclaimer and consistent safety framing. In a spiritual-guidance skill, users may follow these instructions literally, increasing burn, fire, or injury risk despite occasional inline cautions.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.