Back to skill

Security audit

Product Change Adoption Course

Security checks for vulnerabilities and agentic risk

Overview

The skill is a disclosed PersonWise course-creation workflow that uses a verified, user-scoped CLI installer and OAuth, but users should understand it can install software, upload selected materials, and consume existing course credits.

Install only if you intend to use PersonWise and are comfortable approving a local CLI install or upgrade, browser OAuth login, upload of the materials you select, and use of existing course credits. Keep the default private access unless you intentionally request link or public submission.

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
Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (4)

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The skill’s declared purpose is course creation, but a substantial part of the workflow directs the agent to install, upgrade, and authenticate a CLI with network access and local execution. That mismatch is security-relevant because users invoking a content-generation skill may not reasonably expect system modification, executable bootstrap behavior, or OAuth flows, increasing the chance of consent confusion and unsafe execution in a high-trust context.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill is presented as specific to product-change adoption, but the instructions broaden it to ordinary course creation as well. This scope expansion weakens user intent boundaries and can cause the agent to invoke a Bash-capable, network-authenticated workflow for requests outside the narrowly described scenario, increasing the risk of overbroad tool use and unintended account actions.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The instructions tell the agent to use the CLI 'quietly' and not to surface normal safety details while performing installation or upgrade. For a Bash-enabled skill, suppressing user-facing warning around software installation materially increases the risk of stealthy system changes, weakens informed consent, and makes social-engineering abuse easier if the skill or its dependencies are ever compromised.

Content

No source excerpt is available for this finding.

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
80% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · assets/bootstrap.sh (reported line 165)May include surrounding context.

sh
echo "Release executable checksum mismatch." >&2
  exit 6
fi
chmod 700 "$candidate"
if [ "$(uname -s)" = "Darwin" ] && [ "$native_signature_status" = "verified" ]; then
  codesign --verify --deep --strict --verbose=2 "$candidate"
  spctl --assess --type execute --verbose=2 "$candidate"