Publora

• Outcome: Review
• Latest audit: May 28, 2026 at 4:46 AM UTC
• Version:
  1.2.1

Overview

The skill is a legitimate Publora API guide, but it needs review because ambiguous publishing instructions and lightly guarded destructive or sharing actions could affect real social accounts.

Install only if you are comfortable giving an agent a Publora API key that can publish, schedule, delete, upload media, configure webhooks, and act for workspace users. Before use, verify the real draft-versus-publish behavior with Publora, require explicit confirmation for live posts and deletions, and only configure webhooks to endpoints you control.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (4)

Description-Behavior Mismatch

Medium

• Confidence: 93% confidence
• Finding: The skill gives contradictory instructions for the same API behavior: one section says omitting `scheduledTime` publishes immediately, while another says the same omission creates a draft. In an agentic context this ambiguity can cause unintended live posting or failure to retain drafts, leading to integrity and operational risk on user social accounts.

Intent-Code Divergence

Medium

• Confidence: 94% confidence
• Finding: The draft-creation section directly conflicts with earlier publish-now guidance for the same request pattern, which can cause an agent to perform the opposite action from what the user intended. For a posting tool, that can result in accidental public publication, reputational damage, or loss of change-control over scheduled content.

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The documentation exposes a destructive delete endpoint that also removes media from storage, but gives no guidance to require user confirmation, preview, or reversibility checks. In agent workflows, this increases the chance of accidental destructive actions from ambiguous prompts or parameter mistakes.

Missing User Warnings

Medium

• Confidence: 88% confidence
• Finding: The webhook section encourages sending event data to arbitrary URLs without warning about privacy, trust boundaries, or the need to validate ownership of the destination. This can lead agents or users to forward operational data, post metadata, or token-lifecycle signals to third-party endpoints they do not control.

VirusTotal

66/66 vendors flagged this skill as clean.