Paperclip Resilience

Security checks across malware telemetry and agentic risk

Overview

This appears to be a legitimate Paperclip orchestration skill, but it needs review because it can automatically send task/session data to Paperclip APIs and arbitrary webhooks with unclear user controls.

Review this skill before installing in any workspace with secrets, customer data, private code, or sensitive business context. Use it only with trusted Paperclip instances and webhook endpoints, avoid placing secrets in tasks, and configure explicit approvals, retry limits, logging, and redaction where available.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (3)

Missing User Warnings

Medium

Confidence: 80% confidence
Finding: The documentation describes an automated recovery process that re-invokes agents after failed runs without clearly emphasizing the operational and security consequences of repeated autonomous retries. In an agent orchestration context, automatic re-runs can amplify harmful actions, repeat expensive operations, or retrigger failure loops if the original task or environment was unsafe.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The blocker-routing documentation permits sending session-derived blocker data to arbitrary webhooks but does not warn that transcripts or metadata may contain sensitive information such as credentials, internal paths, or business context. In a production orchestration skill, this omission can lead users to exfiltrate sensitive agent output to third-party endpoints without understanding the privacy implications.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The function sends the full task text to an external Paperclip API as part of the auto-created issue description, but there is no user-facing notice, consent check, or data-classification guard before transmission. In an agent-orchestration context, tasks can easily contain secrets, proprietary code, internal URLs, or customer data, so automatic export of raw task content creates a real confidentiality and compliance risk.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal