Relationship Chat Analysis

• Outcome: Pass
• Latest audit: Sep 16, 2026 · 9:04 AM UTC
• Version: 0.1.1

Overview

This skill is a clearly scoped relationship chat analysis helper with strong privacy, consent, and data-minimization instructions.

Install only if you are comfortable giving the agent specific relationship chat records to analyze. Prefer redacting unrelated names, phone numbers, addresses, and third-party details before use, and review any planned output paths before allowing files to be written.

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

SkillSpector

Vulnerability Patterns

• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Memory Poisoning: Persistent Context Injection, Context Window Stuffing, Memory Manipulation
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (3)

Memory Manipulation

High

• Category: Memory Poisoning
• Confidence: 80% confidence

Finding

Skill manipulates agent memory, state, or stored context. Memory corruption can alter personality, override safety rules, or cause unpredictable behavior.

Content

Scanner excerpt · references/instructions.md (reported line 164)May include surrounding context.

md

Pattern confidence:

- `high`: repeated in 3+ episodes, clear message evidence, low missing-context risk
- `medium`: repeated in 2+ episodes or strong single episode with clear context
- `low`: plausible but limited, ambiguous, or context-dependent

## Stage 6: Cross-Episode Synthesis

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 · references/blindspot-prompt.md (reported line 24)May include surrounding context.

md

- Asking for reassurance in a way that lands as accusation
- Treating silence as proof when multiple interpretations remain possible
- Accepting apology words without checking behavior change
- Over-explaining while the other person disengages
- Ignoring repeated boundary signals

Missing User Warnings

Medium

• Category: Not specified by scanner
• Confidence: 90% confidence

Finding

This prompt is designed to process intimate relationship chats and explicitly extract safety-related signals such as coercion, self-harm, harassment, and sexual pressure, but it provides no privacy notice, data-minimization guidance, or handling constraints for highly sensitive personal data. In this skill context, users may paste full chat exports containing third-party messages and crisis content, so omission of privacy and consent safeguards increases the risk of over-collection, inappropriate retention, and unsafe handling of sensitive interpersonal information.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.