tech recruiter pro

• Outcome: Review
• Latest audit: May 29, 2026 at 9:22 PM UTC
• Version:
  1.0.2

Overview

This recruiting skill is mostly coherent, but it includes under-scoped cross-platform candidate profiling and explicit guidance to continue after anti-bot challenges using a proxy.

Review before installing. Use this only for lawful, job-relevant recruiting with approved data sources, official APIs where available, and clear candidate notice or opt-out where required. Do not follow the proxy/captcha workaround; treat anti-bot challenges as a stop condition. Store any exported reports or Feishu records in restricted locations, minimize collected personal data, and avoid committing API tokens or config files containing secrets.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection

Findings (22)

Intent-Code Divergence

Medium

• Confidence: 95% confidence
• Finding: The document makes an absolute security claim that the skill will not send data to unauthorized third parties, while elsewhere it explicitly introduces integrations with external services such as GitHub, LinkedIn, Twitter/X, and Feishu. This inconsistency can mislead users and reviewers about actual data flows, causing them to underestimate privacy and compliance risk when candidate data may be transmitted off-host under some configurations.

Intent-Code Divergence

Medium

• Confidence: 93% confidence
• Finding: The file asserts the skill will not violate robots.txt, yet elsewhere describes HTML parsing, CSS selectors, and scraping-style collection on platforms that may lack official APIs or prohibit automated access. Such overbroad assurances can encourage unsafe deployment and obscure legal/compliance exposure, especially if operators assume the implementation already enforces per-site restrictions when it does not.

Intent-Code Divergence

High

• Confidence: 98% confidence
• Finding: The skill explicitly instructs users to "pause 5 minutes, use proxy IP" when Google Scholar presents a captcha, which is anti-abuse evasion guidance. That directly contradicts the later security declaration that the skill will not bypass platform authentication or platform protections, and it can facilitate scraping in ways that violate site terms or trigger account/IP enforcement.

Intent-Code Divergence

Medium

• Confidence: 94% confidence
• Finding: The skill claims it will not violate robots.txt or bypass protections, but its documented scraping workflow includes behavior intended to continue collection after anti-bot challenges. This inconsistency is dangerous because users may rely on the safety declaration while following instructions that encourage non-compliant scraping and evasion of platform safeguards.

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The release notes advertise cross-platform candidate search, profile analysis, and personalized outreach, but provide no warning about privacy, consent, scraping restrictions, or responsible use of personal data. In a recruiting context, this omission can normalize collecting and using candidate information from multiple external services without clear safeguards, increasing the risk of privacy violations, policy breaches, and harmful unsolicited outreach.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The configuration section lists several API tokens but does not instruct users to store them securely or avoid hardcoding and sharing them. This can lead to credential leakage in source control, logs, or documentation, enabling unauthorized access to third-party accounts and APIs.

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: This preview describes cross-platform candidate sourcing, profiling, scoring, and automated outreach using public data, but it does not present a concrete in-product consent, disclosure, retention, or lawful-basis workflow. Even if the data is public, aggregating it across platforms and using it for recruiting decisions can create privacy, compliance, and fairness risks, especially when exported or used for automated follow-up.

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: The guide encourages exporting candidate lists and recruitment reports, which likely contain personal data, but gives no guidance on consent, access control, retention, or secure handling. In a recruiting context this increases the risk of unauthorized disclosure, over-sharing, or noncompliant processing of candidate information.

Missing User Warnings

Low

• Confidence: 88% confidence
• Finding: The workflow describes analyzing named candidates, reviewing publication and GitHub profiles, and generating personalized outreach without any notice that personal profile data will be processed. In a recruiting skill, this omission is material because users may input or aggregate personal data from multiple sources without understanding privacy expectations, lawful basis, or minimization requirements.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: This section explicitly describes building a unified candidate profile by aggregating personal and professional data from multiple platforms, but provides no guidance on consent, lawful basis, data minimization, retention, or access controls. In a recruiting skill, this increases the risk of privacy-invasive profiling, noncompliant processing of personal data, and downstream misuse of sensitive inferences.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The examples show searching for named individuals, usernames, and employer affiliation across multiple services for recruiting evaluation without any privacy warning or anti-doxxing safeguards. That makes the workflow more dangerous because it normalizes targeted profiling of identifiable people and could facilitate intrusive vetting or collection of excessive personal data.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The guide explicitly encourages building a unified candidate profile by aggregating data from multiple platforms, including professional, academic, and community sources, but provides no privacy, consent, retention, or lawful-basis guidance. In a recruiting context, this can enable excessive profiling, collection of sensitive personal data, and misuse of cross-platform identity correlation, creating compliance and privacy risk even if no exploit code is present.

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The API configuration examples include environment variables for secrets and bot tokens but do not warn about secure secret handling, least privilege, rotation, or avoiding commits/logging. This omission can lead users to expose credentials in shell history, screenshots, repositories, CI logs, or shared environments, which may allow unauthorized access to third-party accounts and data.

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The guide explicitly shows storing a live Twitter Bearer Token in config.ini, which encourages placing long-lived credentials in a file that may be committed to source control, copied into logs, or exposed through backups and screenshots. In this context, the skill is an operational guide for API usage, so the example is more dangerous because readers are likely to follow it verbatim and handle real credentials insecurely.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The report export writes recruiter-collected personal data, including names, employer, position, research areas, citation metrics, and profile links, to a local markdown file without consent controls, minimization, access restrictions, or even an explicit warning. In a recruiting context this is sensitive professional profiling data, and silent disk persistence increases the chance of unintended disclosure through shared workspaces, backups, or temporary directories.

Vague Triggers

Medium

• Confidence: 83% confidence
• Finding: The example trigger '搜索 RLHF 方向候选人，目标公司 Moonshot/DeepSeek/DeepMind' is broad enough that ordinary recruiting requests could invoke this skill without clearly signaling that it will perform multi-platform people search and candidate profiling. In a skill that aggregates data from GitHub, Google Scholar, social platforms, and pipeline tooling, ambiguous activation increases the chance of unintended collection and processing of personal data.

Missing User Warnings

Medium

• Confidence: 90% confidence
• Finding: The manifest advertises candidate search, smart profile analysis, personalized email generation, and pipeline management, but the top-level description does not present a clear privacy notice or consent/data-handling warning to the user. Because the skill is explicitly designed to collect, correlate, and operationalize information about identifiable individuals across multiple platforms, insufficient disclosure raises privacy, compliance, and misuse risks.

Unpinned Dependencies

Low

• Category: Supply Chain

Content

requests>=2.28.0
beautifulsoup4>=4.11.0
lxml>=4.9.0

• Confidence: 87% confidence
• Finding: requests>=2.28.0

Unpinned Dependencies

Low

• Category: Supply Chain

Content

requests>=2.28.0
beautifulsoup4>=4.11.0
lxml>=4.9.0

# Optional dependencies

• Confidence: 87% confidence
• Finding: beautifulsoup4>=4.11.0

Unpinned Dependencies

Low

• Category: Supply Chain

Content

requests>=2.28.0
beautifulsoup4>=4.11.0
lxml>=4.9.0

# Optional dependencies
# github-api  # For enhanced GitHub search

• Confidence: 90% confidence
• Finding: lxml>=4.9.0

Known Vulnerable Dependency: requests — 10 advisory(ies): CVE-2014-1830 (Exposure of Sensitive Information to an Unauthorized Actor in Requests); CVE-2024-47081 (Requests vulnerable to .netrc credentials leak via malicious URLs); CVE-2024-35195 (Requests `Session` object does not verify requests after making first request wi) +7 more

High

• Category: Supply Chain
• Confidence: 94% confidence
• Finding: requests

Known Vulnerable Dependency: lxml — 10 advisory(ies): CVE-2021-43818 (lxml's HTML Cleaner allows crafted and SVG embedded scripts to pass through); CVE-2014-3146 (lxml Cross-site Scripting Via Control Characters); CVE-2021-28957 (lxml vulnerable to Cross-Site Scripting ) +7 more

High

• Category: Supply Chain
• Confidence: 92% confidence
• Finding: lxml

VirusTotal

65/65 vendors flagged this skill as clean.