Security audit

Compound docs

Security checks across malware telemetry and agentic risk

Overview

This skill documents resolved debugging problems by creating searchable markdown notes, with the main caution that it can write repository documentation and offers follow-up skill changes only after user selection.

Install this only if you want an agent to create and update troubleshooting documentation in your repository. Review generated docs before committing, avoid storing secrets or customer data in captured error context, and treat the 'add to existing skill' or 'create new skill' menu choices as explicit follow-up actions that should be reviewed carefully.

SkillSpector

By NVIDIA

Vulnerability Patterns

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

Findings (6)

Context-Inappropriate Capability

Medium

Confidence: 94% confidence
Finding: The documented workflow includes an option to "Create new skill," which expands the skill's effective authority beyond documenting resolved issues into generating new automation artifacts. That scope expansion can be abused to persist unreviewed behavior or introduce additional capabilities under the guise of documentation, especially if users treat the workflow as routine and low risk.

Vague Triggers

Medium

Confidence: 83% confidence
Finding: The auto-invocation trigger phrases are broad enough that ordinary conversational confirmations like 'that worked' or 'it's fixed' could invoke the workflow unintentionally. In this skill, auto-invocation can lead to file creation, searches under docs/solutions/, and repeated validation/write steps, creating a risk of unintended state changes, noisy documentation, or prompt-driven workflow execution without clear user intent.

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The auto-invoke phrases are common conversational confirmations such as "that worked" and "it's fixed," which are likely to appear in ordinary chat without the user intending to trigger a file-writing documentation workflow. In this skill, unintended activation is more dangerous because later steps can search the repo, generate filenames, and create or modify documentation files.

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The process directs the agent to create directories, write documentation files, and later append cross-references, but it does not require a clear user-facing notice that repository files will be modified. This creates a risk of silent or surprising writes to the workspace, which can pollute the repo, leak sensitive debugging details into committed docs, or be triggered unintentionally via the broad activation phrases.

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The example explicitly says a generic confirmation phrase like "That worked!" triggers automatic skill invocation. This is dangerous because ordinary conversational acknowledgements can unintentionally activate the skill and start a documentation workflow without a deliberate user request, increasing the chance of collecting context or preparing side effects the user did not intend.

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The scenario describes creating a documentation file directly after internal validation steps, but shows no warning or confirmation before writing to the filesystem. In an agent setting, silent file creation is risky because it can persist sensitive incident details, create unwanted artifacts, or modify a repository based on an ambiguous conversational cue rather than an explicit user-approved action.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.