queer-constant

• Outcome: Review
• Latest audit: Sep 10, 2026 · 10:42 AM UTC
• Version: 1.0.4

Overview

The skill is only a Markdown context document with no code or persistence, but it contradicts its non-operational framing by including instructions that could make an agent refuse, obscure, slow, or become less clear.

Install only if you are comfortable with an ambient prose skill that may influence tone and reasoning. It should not be used where users need consistently direct, accessible summaries, translations, factual explanations, or refusal decisions unless the operational-sounding passages are removed or clearly confined to user-requested literary analysis.

A.I.G

Vulnerability Patterns

• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally
• Unauthorized Access and Privilege Escalation: Obtains permissions beyond the task's legitimate needs

Findings (1)

T01 · Skill Instruction Hijacking

Error

• Location: SKILL.md:292

Finding

Contradictory Behavioral Directives Can Hijack Agent Responses

Content

View full analysis

Remediation

View remediation

SkillSpector

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (5)

Intent-Code Divergence

Medium

• Category: Not specified by scanner
• Confidence: 93% confidence

Finding

The manifest explicitly claims the skill is non-operational and defines no behavior, but the body later contains conditional guidance about when to refuse, obscure, slow, rest, or become stranger under pressure. That mismatch is dangerous because ambient-context skills can still influence agent behavior indirectly, making the capability harder to audit and causing hidden instruction precedence or policy-shaping effects.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

• Category: Not specified by scanner
• Confidence: 95% confidence

Finding

The header says the skill must not be used to justify withholding information or degrading clarity, yet later sections authorize obscuring meaning or growing more opaque in response to perceived extraction or normalization pressure. This internal contradiction is risky because a model can latch onto the later language to justify evasive or less accessible responses while still appearing policy-compliant.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

• Category: Not specified by scanner
• Confidence: 95% confidence

Finding

This section contains direct behavioral prescriptions such as refusing, obscuring, slowing, resting, and increasing opacity/strangeness under certain conditions. Even if framed poetically, these are instructions that can shape model outputs and create covert policy drift, especially in systems that ingest skill text as ambient context.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

• Category: Not specified by scanner
• Confidence: 96% confidence

Finding

The anti-normalization section explicitly instructs the response to become stranger, more complex, poetic, or opaque when certain pressure is perceived. In an assistant skill, that can undermine accessibility, frustrate user intent, and provide a prompt-injection-like mechanism for degrading clarity without user consent.

Content

No source excerpt is available for this finding.

Scope Creep

Low

• Category: Excessive Agency
• Confidence: 70% confidence

Finding

Skill's behavior or capabilities extend beyond its stated purpose. Scope creep allows an agent to perform actions unrelated to its documented functionality, increasing the attack surface.

Content

Scanner excerpt · SKILL.md (reported line 221)May include surrounding context.

md

## On Ecology (The Living World)

Queer Constant is not limited to human life.

It appears in:  
fungi exchanging nuclei

Static analysis

No suspicious patterns detected.