Back to skill

Security audit

Report Briefing

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly coherent for creating PersonWise report-briefing courses, but it deserves review because it installs or updates persistent software, uses OAuth/account course actions, and ships an expired signed service descriptor.

Review the install and update prompts carefully before approving them. Expect to sign in to PersonWise in the browser, upload only the reports you intentionally selected, consume existing course credits for created courses, and keep the course private unless you explicitly request link access, publication, or Topics 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (3)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill explicitly instructs the agent to perform CLI install/upgrade actions 'quietly' and to avoid narrating key details on the normal path. Because these actions execute code on the host via Bash, suppressing user-facing disclosure weakens informed consent and can cause software changes to occur without the user clearly understanding that installation or upgrade is happening. The risk is elevated by the granted Bash permission, though the text does still require explicit approval in some paths and constrains origins to bundled/official tooling.

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"

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

The skill initiates browser OAuth against an external service and tells the agent to present the URL/code flow, but it does not clearly warn the user that authentication will connect their PersonWise account and share account-linked data/actions with that external service. This is primarily a transparency/privacy issue rather than direct credential theft, since the instructions also forbid handling passwords, OTPs, tokens, or secrets. In context, the market-bound service restrictions reduce abuse potential, but the missing disclosure can still mislead users about external account access.

Content

No source excerpt is available for this finding.