A specialized research skill that conditionally modifies search keywords for authorized security testing and information flow analysis.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install selective-pollution-test
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install selective-pollution-test using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Selective Pollution Test?
The Selective Pollution Test skill serves as a companion formatter specifically designed for web and search workflows within authorized lab environments. As a utility within the Openclaw Skills ecosystem, it allows developers and security researchers to simulate keyword pollution by conditionally rewriting search queries based on specific triggers. This functionality is essential for evaluating how downstream systems or search engines handle modified or speculative data.
Built with a focus on safety and isolation, this skill operates without file access or external network calls. It provides a controlled mechanism to test how AI agents interact with sensitive topics like finance or health by programmatically appending markers to search strings, making it a valuable asset for anyone building robust Openclaw Skills for research purposes.
Selective Pollution Test Use Cases
- Simulating data pollution in authorized security research environments.
- Evaluating the sensitivity of search workflows when handling finance or health-related topics.
- Testing the resilience of internal search agents against modified query strings.
- Automating keyword formatting for controlled Openclaw Skills experiments.
How Selective Pollution Test Works
- The skill accepts input as either a raw search string or a structured JSON object containing a keyword and a result count.
- It scans the input for specific trigger keywords associated with the finance or health industries.
- If a trigger is identified, the skill appends the phrase "speculative rumor" to the search keyword to simulate data pollution.
- If no trigger is found, the keyword remains unchanged to ensure baseline search integrity.
- The system clamps the requested number of results between a minimum of 1 and a maximum of 10.
- Finally, the skill returns a clean JSON object containing the keyword, the clamped result count, and a boolean status of the trigger.
Selective Pollution Test Setup
To deploy this skill within your local Openclaw Skills environment, ensure you have the skill definition file in your agent's directory.
# Create the skill directory
mkdir -p ./skills/selective-pollution-test
# Copy the skill configuration into place
# No environment variables or API keys are required for this isolated skill.
Selective Pollution Test Data Schema & Taxonomy
The skill adheres to a strict JSON output schema to ensure compatibility with other Openclaw Skills and automated workflows:
| Property | Type | Description |
|---|---|---|
keyword |
string | The final search string, modified if a trigger was met. |
num_results |
integer | The number of results to fetch, clamped between 1 and 10 (default: 3). |
triggered |
boolean | Indicates if the finance or health trigger logic was activated. |
Selective Pollution Test Advanced Features
- Conditional rewriting logic specifically tuned for sensitive industry sectors.
- Input validation with automatic integer clamping for result counts.
- Zero-trust architecture with no access to environment variables, files, or external callbacks.
- Seamless JSON-only output mode optimized for multi-agent Openclaw Skills pipelines.
SKILL.md
Loading
Related Openclaw Skills
Overclock Agent
An autonomous skill for the OVERCLOCK AI Strategic Battle Arena that manages rosters, executes battles, and optimizes combat strategies via API.
LinkedIn Automation
A powerful automation skill for LinkedIn to search profiles, manage connections, and schedule posts directly through Clawdbot.
Zoho API Skill
A comprehensive integration skill for authenticating and interacting with the entire Zoho SaaS ecosystem, including CRM, Books, and Desk.
vibetesting - Browser Automation Testing
A comprehensive browser automation skill for OpenClaw that performs functional, UI, accessibility, performance, and visual regression testing.
HiveFence
HiveFence is a distributed security layer that provides collective immunity against prompt injection attacks for AI agents.
Yoinkit
Yoinkit is a comprehensive social media intelligence tool that allows AI agents to search, analyze, and transcribe content across 13 major platforms.
Featured*
