Security audit

PagerDuty On-Call

Security checks across malware telemetry and agentic risk

Overview

This PagerDuty skill does what it claims, but it gives an agent live incident-management power without clear approval safeguards.

Review before installing. Use a least-privileged PagerDuty token limited to the services this agent should manage, protect API and integration keys from logs or chat transcripts, and require explicit human approval before acknowledge, resolve, trigger, snooze, reassign, note, or direct curl API actions.

SkillSpector

By NVIDIA

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration

Findings (5)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 95% confidence
Finding: The skill clearly requires environment access for `PAGERDUTY_API_KEY` and network access to the PagerDuty API, yet no permissions are declared. Missing capability declarations weaken platform trust boundaries because users and policy engines are not explicitly informed that the skill can access secrets and perform outbound authenticated API actions.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The documentation encourages operationally destructive actions such as acknowledging, resolving, triggering, snoozing, and reassigning incidents, but provides no warning, confirmation guidance, or change-control expectations. In an agent context, this increases the risk of unintended production-impacting actions being taken automatically or on ambiguous user requests.

Missing User Warnings

Medium

Confidence: 84% confidence
Finding: The acknowledge and resolve commands perform state-changing PagerDuty actions immediately with no confirmation, dry-run mode, or explicit user-warning gate. In an agent setting, this increases the risk of accidental incident mutation, which can suppress alerts or prematurely resolve active incidents and disrupt operations.

Missing User Warnings

Medium

Confidence: 87% confidence
Finding: The trigger command sends incident data to PagerDuty and creates a new incident without any advance disclosure or confirmation. In an agent workflow, this can cause false alerts, notification storms, and operational disruption if invoked from ambiguous or manipulated instructions.

Missing User Warnings

Medium

Confidence: 86% confidence
Finding: Adding notes, snoozing incidents, and reassigning ownership are all state-changing operations that execute immediately without confirmation or safety interlocks. In agent-driven environments, these actions can alter incident handling, hide urgency, or redirect responsibility based on mistaken or adversarial prompts.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.