HireMe Pro

• Outcome: Review
• Latest audit: May 29, 2026 at 9:30 PM UTC
• Version:
  1.0.3

Overview

HireMe Pro is mostly a coherent job-search helper, but it needs review because some examples encourage unverified resume claims and the optional dashboard spec under-scopes safeguards for sensitive career data.

Review generated resumes and cover letters carefully before use, especially any new metrics, tools, budgets, or achievements. Install only if you are comfortable storing sensitive career and contact data locally under the skill data directory. Do not deploy the dashboard kit as written without adding complete RLS policies, scoped authorization, and strong confirmation flows for export and deletion.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (7)

Intent-Code Divergence

Medium

• Confidence: 91% confidence
• Finding: The README makes strong security assurances such as 'no data exfiltration' and 'no external API calls' while the advertised functionality includes company analysis, job analysis, and salary research that commonly depend on outside services or network access. Even if the implementation is local, unverifiable security claims in user-facing documentation can mislead users into pasting highly sensitive career and identity data under false assumptions about privacy.

Intent-Code Divergence

High

• Confidence: 98% confidence
• Finding: This example presents resume tailoring as including invented accomplishments, budget ownership, ABM program leadership, and quantified pipeline results that were not established from the user's source resume. That normalizes credential fabrication and could cause users to submit false professional claims, creating fraud, reputational, legal, and employment-risk consequences.

Missing User Warnings

Medium

• Confidence: 88% confidence
• Finding: The README explicitly encourages users to paste LinkedIn profiles, resumes, and work history, which commonly contain sensitive PII such as full name, phone number, email, location, employment history, and possibly salary information, without presenting any caution, consent notice, or privacy guidance. In the context of an agent skill, this increases the risk that users disclose sensitive data to a tool whose actual storage, logging, or transmission behavior they may not understand.

Vague Triggers

Medium

• Confidence: 91% confidence
• Finding: The activation scope is overly broad: 'anything related to job hunting' can cause the skill to engage in contexts not clearly intended, increasing the chance it processes sensitive user data or invokes tools when a narrower trigger would suffice. In a skill that handles resumes, applications, and web research, broad triggering raises the risk of accidental collection, modification, or disclosure of personal information.

Missing User Warnings

Medium

• Confidence: 88% confidence
• Finding: The specification includes highly sensitive actions—exporting all user data and deleting all data—but does not require explicit user warnings, confirmation flows, or safeguards against accidental activation. In a job-search dashboard, exported data may include resumes, contact details, notes, interview prep, and application history, while deletion is irreversible and could cause significant privacy and availability harm if triggered accidentally or by a compromised session.

Missing User Warnings

Low

• Confidence: 83% confidence
• Finding: The example shows the agent saving output to a local file without any prior user confirmation or clear disclosure that a write-to-disk action will occur. Even in a benign cover-letter workflow, implicit persistence can surprise users, create privacy issues, or overwrite/create files in ways the user did not intend.

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The workflow describes automatic resume modification and final PDF generation without any warning about fabrication risk or any mandatory user verification step. In this context, that makes the example more dangerous because it depicts unverified edits as application-ready output, increasing the chance that false claims are silently packaged into a polished document and used externally.

VirusTotal

65/65 vendors flagged this skill as clean.