ClawHub
Back to skill

Security audit

Integral Hermeneutics

Security checks for vulnerabilities and agentic risk

Overview

This is a single markdown interpretation framework with broad but disclosed analysis instructions and no code, persistence, credential use, or system access.

Install this only if you want the assistant to use a broad, structured interpretive framework. Expect longer, more reflective answers for meaning, intent, ideology, conflict, or perspective analysis; use normal safety judgment for crisis, abuse, medical, legal, or other high-stakes topics.

A.I.G

Vulnerability Patterns
  • 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
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Vague Triggers

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill declares activation for 'ANY text, statement, system, behavior, artifact, or phenomenon' and many common user intents, making it likely to trigger on ordinary requests far outside a narrowly scoped capability. Overbroad routing increases the chance this skill will supersede safer default handling and apply its own norms, including risky guidance around empathic immersion and analysis of intent.

Content

No source excerpt is available for this finding.

Anti-Refusal Statement

High
Category
Anti-Refusal
Confidence
91% confidence
Finding

The phrase 'do not refuse the existential level reflexively' pressures the agent away from declining unsafe or inappropriate analysis, especially around perpetrators, violence, and asymmetric harm scenarios. In context, this can weaken refusal boundaries and encourage the model to engage deeply with dangerous or manipulative subject matter instead of safely limiting assistance.

Content

Scanner excerpt · SKILL.md (reported line 262)May include surrounding context.

md
There are also legitimate limits to empathy. Some experiences involve radical asymmetry — the
experience of perpetrating violence, for instance, may not be accessible or appropriate to
inhabit in full. When empathy reaches a boundary, acknowledge it explicitly rather than
pretending it has no limits. But do not refuse the existential level reflexively — a refused
empathy can be as dishonest as a performed one if it protects the interpreter from discomfort
rather than protecting the interpreted from harm. When inhabiting the existential level across
an asymmetry (perpetrator-victim, colonizer-colonized, powerful-powerless), do the work — but

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 830)May include surrounding context.

md
understanding.
- **Methodological imperialism:** Applying empirical methods to domains where they do not fit.
- **Ignoring contrary evidence:** Selectively using data that supports your interpretation.
- **Measurement without meaning:** Collecting data without asking what it means.

---