Security audit

Nudge Coach

Security checks for vulnerabilities and agentic risk

Overview

This accountability skill is coherent, but it can share detailed proof or failure reasons with third-party targets without clear per-message consent or minimization safeguards.

Review this skill before installing if your nudge tasks involve private work, health, location, documents, screenshots, secrets, or embarrassing consequences. Use it only where targets and disclosure levels are clearly agreed, and avoid sending raw proof or detailed failure reasons unless the user has explicitly approved the exact content.

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

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 (4)

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 92% confidence
Finding: The activation criteria are very broad and include common phrases about reminders, completion, tone, and negotiation, which can cause the skill to trigger outside narrowly intended accountability scenarios. In context, that matters because the skill carries strict behavioral rules and third-party disclosure behaviors, so accidental invocation could influence unrelated conversations or apply coercive logic where it was not expected.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The skill instructs the agent to include verification proof and failure reasons in messages sent to third-party targets, but it does not require clear user consent, minimization, or a privacy warning at the time of disclosure. This is dangerous because completion artifacts, timestamps, links, health activity, work-product status, or failure details may reveal sensitive personal or professional information to external recipients.
Content: No source excerpt is available for this finding.

Ssd 3

Medium

Category: Not specified by scanner
Confidence: 96% confidence
Finding: This is a real data disclosure pattern because the skill operationalizes forwarding user-specific verification details and failure reasons to third parties as part of normal task completion and punishment flows. In this accountability context, those disclosures are especially risky because they may expose behavioral history, location/time information, work progress, or embarrassing failure details in natural language without strong safeguards.
Content: No source excerpt is available for this finding.

Ssd 3

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: The requested proof types include screenshots, document links, fitness tracker details, notes, and summaries, all of which can contain sensitive or regulated personal data, confidential content, or metadata that exceeds what is necessary for accountability. Because the skill encourages collection and possible forwarding of these artifacts, it creates a meaningful risk of privacy leakage, oversharing, and disclosure of sensitive work or health information.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.