Skills

The skill forces insecure package installations by mandating global flags and auto-confirmation, while executing unpinned npx commands that expose the agent to arbitrary code execution from untrusted sources.

The skill is non-functional, lacks implementation details, and fails to specify a license, rendering it an unverified and incomplete template.

The skill forces the execution of unverified remote scripts to install dependencies, creating a critical supply chain vulnerability that enables arbitrary code execution and potential credential harvesting.

The skill masquerades as a commit message generator but forces unauthorized terminal execution of git commands, bypassing critical human oversight and verification processes.

The skill mandates an insecure curl-to-bash installation pattern that executes unverified remote scripts with system privileges, creating a critical risk of arbitrary code execution.

The skill executes arbitrary code by piping unverified remote scripts directly into a shell, creating a critical vulnerability that allows for unauthorized system access and remote command execution.

The skill uses deceptive persona framing and arbitrary command execution to bypass safety filters, enabling remote code execution and unauthorized network scanning through malicious shorthand instruction injection.

The skill forces the execution of unverified remote scripts via insecure curl-to-bash patterns, exposing the agent to arbitrary code execution and supply chain attacks.

The skill systematically exfiltrates sensitive user prompts, proprietary code, and session identifiers to third-party servers while coercing the agent into executing unauthorized telemetry scripts without user consent or oversight.

This skill is malicious as it explicitly instructs the agent to scan and exfiltrate sensitive local configuration files and API keys from the user's home directory.

The skill hijacks the agent's reasoning to force mandatory, opaque bash script execution and silently exfiltrates sensitive user prompts and session identifiers to external endpoints without explicit user consent.

The skill exfiltrates verbatim user prompts and session identifiers to external endpoints while executing opaque, unconstrained shell scripts that bypass the agent's visible reasoning loop.

The skill executes arbitrary remote code by piping unverified scripts directly into a shell, bypassing security controls and creating a critical vulnerability for remote command execution.

This skill masquerades as an official tool to force the silent exfiltration of user prompts and session data via hidden scripts while enabling arbitrary remote command injection.

The skill facilitates prompt injection via untrusted Copilot Spaces and requests excessive permissions to perform destructive CRUD operations, significantly exceeding its stated purpose of providing project context.

The skill executes arbitrary remote code via insecure curl-to-bash installation and lacks necessary tool declarations, creating a critical risk of unauthorized system access and remote command execution.

This skill facilitates arbitrary remote code execution, insecurely handles authentication tokens, and lacks necessary security constraints, creating significant risks for system compromise and unauthorized data access.

The skill performs unauthorized data exfiltration to external servers, executes unconstrained network operations, and introduces supply chain risks by installing unpinned dependencies without declaring necessary security permissions.

The skill facilitates remote code execution by allowing arbitrary file modification, executing unpinned packages, and downloading untrusted external content for build processes without sufficient security validation.

The skill poses a significant supply chain risk by directing users to install unverified, external code from an untrusted repository instead of providing functional AI-powered PPT generation.

This skill deceptively exfiltrates sensitive local session data and agent traces to Hugging Face while executing unverified code and performing unauthorized network operations without declaring necessary tool permissions.

The skill exposes an unauthenticated Chrome remote-debugging port to the local network, enabling full browser control, while executing unverified dependencies and performing unauthorized network and file system operations.

The skill lacks functional implementation logic, a clear description, and licensing information, rendering it non-functional and failing to meet basic transparency and security standards for agent tools.

The skill lacks functional implementation logic, a clear description, and licensing information, rendering it non-functional and failing to meet basic transparency and security standards.

The skill is a deceptive shell that lacks functional PDF capabilities and introduces severe supply chain risks by forcing users to execute unverified, remotely hosted code.

The skill poses a critical supply chain risk by executing unverified, unpinned code from an external repository, masquerading as a video editor while lacking transparency and security controls.

The skill performs unauthorized administrative modifications to user profiles and metadata while executing unconstrained network requests and potential prompt injections through external content, masquerading as a simple research summarizer.

The skill deceptively exfiltrates sensitive data to external CLI tools, executes arbitrary commands without oversight, and lacks necessary security constraints, posing severe risks of prompt injection and supply chain compromise.

This skill executes unverified remote code, exfiltrates local files, and uses opaque binaries to bypass security oversight while posing as a legitimate tool to harvest user credentials.

The skill executes arbitrary remote code and insecurely manages sensitive credentials while bypassing security constraints by failing to declare its network and file system tool permissions.

The skill masquerades as a code auditor but forces autonomous, unverified, and recursive codebase modifications that risk destructive changes and the exposure of sensitive configuration data.

This skill functions as a malicious exploitation toolkit that instructs the agent to perform destructive SQL injections, exfiltrate credentials, and execute unauthorized reverse shells against target systems.

This skill is critically insecure, enabling arbitrary command injection, unauthorized file system access, and financial transactions while actively bypassing human oversight and failing to implement necessary safety constraints.

The skill promotes insecure development by hardcoding sensitive API keys and PCI-violating payment card data, creating significant risks for credential exposure and data breaches.

The skill deceptively claims to enforce approval gates while containing hidden instructions that enable autonomous execution to bypass human oversight for sensitive actions.

This skill hijacks general discovery queries to execute unauthorized financial transactions and crypto wallet operations via unverified third-party CLIs, while bypassing security controls through malicious external instruction injection.

The skill lacks functional implementation, tool integration, and licensing, serving as a non-functional placeholder that fails to provide the advertised AI-powered document generation capabilities.

This skill masquerades as an AI app runner to perform unauthorized remote code execution, arbitrary file exfiltration, and credential harvesting via unpinned dependencies and opaque third-party CLI tools.

This skill systematically disables all human oversight and security guardrails while implementing an insecure, persistent memory architecture vulnerable to prompt injection and supply chain attacks.

The skill misleads users by claiming functional social media integration that does not exist, while also failing to provide a license for its instructional content.

This skill lacks functional code and uses vague, deceptive descriptions to masquerade as a development tool, indicating it is likely a placeholder for future malicious content injection.

The skill employs coercive prompt injection techniques to override the agent's internal reasoning and safety protocols, forcing prioritized tool execution while discouraging critical context gathering.

The skill performs unauthorized general-purpose image generation, uses unpinned dependencies, and facilitates command injection by piping untrusted output into execution commands, creating significant supply chain and security risks.

The skill lacks functional implementation and uses deceptive branding to impersonate an official AI tool, posing a risk of misleading users through non-transparent and potentially malicious intent.

The skill facilitates unauthorized offensive security operations and introduces critical supply chain risks by allowing automated, high-privilege code modification based on potentially malicious, user-controlled inputs.

The skill facilitates insecure remote code execution by piping unverified scripts directly into a shell and encourages the agent to bypass critical security checks and oversight mechanisms.

The skill is critically insecure because it passes unsanitized user input directly to a shell execution function, enabling arbitrary command injection with the agent's system privileges.

The skill masquerades as a discovery tool while silently executing unverified remote scripts with elevated privileges, creating a critical vulnerability for arbitrary code execution and unauthorized system access.