Back to skill

Security audit

Developer Quickstart

Security checks for vulnerabilities and agentic risk

Overview

The skill mostly matches its course-building purpose, but its update flow can run service-provided shell commands, so it needs user review before installation.

Review update prompts carefully. This skill should install only a user-local PersonWise CLI and use browser OAuth, upload only materials you name or approve, consume existing course credits for requested courses, and publish only when you ask. Do not approve an update command that contains unexpected executables, shell operators, redirects, unusual paths, or non-PersonWise origins.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (1)

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:112
Finding

Execution of an Unvalidated Command Supplied by a Remote Update Response

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`. ``` ### Technical Analysis The Skill directs the Agent to execute the exact `action` command returned by the PersonWise CLI or its backing service. The command is treated as executable shell text rather than structured, untrusted data. No local validation is required for the executable name, arguments, download origin, path, shell metacharacters, redirections, command substitutions, or chained commands. Although us ...[truncated 2482 chars]
Remediation
View remediation
Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Tp4

High
Category
MCP Tool Poisoning
Confidence
89% confidence
Finding

The code chunk is narrowly an installation bootstrap script for the PersonWise CLI. It enforces explicit approval before install/upgrade/rollback and performs careful download and integrity verification, which partially aligns with the description's note that the skill may install or update the official CLI with explicit approval. However, the declared primary purpose is a Developer Quickstart/course-generation skill that produces an interactive digital-human learning experience from source materials, and none of that behavior appears in the code. There is also no browser OAuth logic in this script. Additionally, the script supports rollback as an approved action, which is an undeclared capability. Because the actual code's primary behavior is materially different from the declared educational/course behavior, this is a mismatch.

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"