Back to skill

Security audit

Internal Change Adoption Course

Security checks for vulnerabilities and agentic risk

Overview

The skill mostly fits its course-creation purpose, but its updater asks the agent to run update commands supplied by the CLI/service with too little local validation.

Review this before installing if you are comfortable letting the skill install or update a local PersonWise CLI, use browser OAuth, upload selected internal documents, consume existing course credits, and potentially run update commands supplied by PersonWise update responses. The main risk is not the course workflow itself, but the under-scoped update command path.

Vulnerability Patterns
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • 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
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Findings (1)

T03 · Remote Payload Retrieval and Execution

Error
Location
SKILL.md:112
Finding

Execution of Server-Controlled Update Commands

Content
View full analysis
` with the directory of this installed Skill (the directory containing this Skill's SKILL.md). Never run `doctor` or a generic capability preflight to check freshness; the `update check` command above is the freshness check. Never ask more than once per component per session, and never substitute another command, flag, origin, or download path for the printed `action`. If the CLI answers `Unknown command` for `personwise update`, the installed CLI predates this Skill's update tooling: ``` ### Technical Analysis The Skill explicitly instructs the Agent to execute an `action` command returned by the PersonWise CLI or its remote service. The returned value is treated as executable shell text rather than structured data mapped to a locally defined, allowlisted operation. No instruction requires the Agent to: - Parse ...[truncated 2587 chars]
Remediation
View remediation
Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
Findings (4)

Tp4

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding

The description frames the skill primarily as a course-creation tool for internal change adoption training, with optional CLI installation/update as a secondary capability requiring explicit approval. The actual code chunk contains only the installer/bootstrap logic for the PersonWise CLI and none of the stated course-related behavior. While the install/update behavior is partially disclosed, the code's primary and sole function is installation management, including an undeclared rollback path. Because the implemented behavior materially differs from the declared main purpose, this is a mismatch.

Content

No source excerpt is available for this finding.

Scope Creep

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The PowerShell bootstrapper performs an external network download and installs an executable, but the skill metadata declares only Bash permission. That mismatch can defeat user/operator expectations and permission-based safety controls, especially because the downloaded binary is then executed or made available locally, expanding the skill's effective capabilities beyond its declared permission model.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The manifest description says the skill should be used for 'organizational change' and 'internal software rollout' without tighter constraints or negative examples. 'Organizational change' is especially broad and can overlap with many general consulting, HR, or training requests, increasing the chance of unintended invocation.

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"