name	lmeterx-web-loadtest
emoji	⚡
description	LMeterX Web Load test tool。When a user provides a webpage/website URL, the script executes immediately to perform: Page Analysis + API Pre-checks + Batch Creation of Load Testing Tasks. No health checks are required.
triggers	["web loadtest","load test this website","analyze and load test this page","load test this web page","帮我压测这个网站","帮我压测这个网页","压测这个页面","压测一下这个网站","对这个网站进行压测","帮我测试一下这个网站的性能"]

Skill: lmeterx-web-loadtest

⚠️ Intent Routing and Execution Rules (Highest Priority — Must Read)

Mandatory Trigger: Whenever a user mentions "website," "webpage," "load testing," or provides a URL that appears to be a website (e.g., https://www.baidu.com), you must—and may only—use this specific Skill.
Prohibition on Unauthorized API Calls: This Skill is authorized to execute local scripts only; you are strictly forbidden from manually constructing HTTP requests—using tools such as curl or requests—to call any LMeterX APIs.
Prohibition on Fabricating Results: You must genuinely execute the provided scripts within a Bash environment and formulate your response to the user based solely on the actual console output (Stdout) generated by the script.

The Only Correct Way to Execute

Upon receiving the user's instruction, open the terminal (Bash/Shell) and strictly execute the following complete command. To prevent authentication failures caused by empty environment variables, you must explicitly prefix the command with LMETERX_AUTH_TOKEN!

python "${SKILL_DIR}/scripts/run.py" --url "<web URL>"

If the user explicitly specifies parameters—such as the concurrency level—within the conversation, execute the task using these custom parameters:

python "${SKILL_DIR}/scripts/run.py" \
  --url "<web URL>" \
  --concurrent-users <concurrent users> \
  --duration <duration> \
  --spawn-rate <spawn rate>

Internal Mechanism Details (Execution Only—No Manual Intervention Required): The script will automatically complete the following three steps: Page Analysis, API Connectivity Pre-check, and Load Testing Task Creation. The script comes with a built-in default token—lmeterx—which is automatically injected into the X-Authorization request header and bound to the agent user.

Guidelines for Presenting Results to the User

Upon completion of execution, please extract the terminal output and respond to the user clearly using the following Markdown format:

Executive Summary: The number of APIs discovered, and the count of pre-flight checks passed vs. failed.
Failure Classification (If Applicable): A brief summary of the reasons for any pre-flight failures (e.g., 401, 404, 5xx, etc.).
Successful Tasks and Report Links: List the task_id for each successfully created task, along with the corresponding URL to view the report: {LMETERX_BASE_URL}/http-results/{task_id}.

Exception Handling

Error Scenario	Output Message
HTTP 401/403	LMeterX token is invalid or expired; please check the `LMETERX_AUTH_TOKEN` configuration.
HTTP 5xx	LMeterX platform service error; please try again later.
Connection Failure	Unable to connect to the LMeterX service; please check your network connection.
Request Timeout	Request timed out; the target page may be too complex.
Concurrency Out of Range	Concurrency level exceeds the valid range [1, 5000].
Invalid URL Format	The URL must begin with `http://` or `https://`.

Skill: lmeterx-web-loadtest

⚠️ Intent Routing and Execution Rules (Highest Priority — Must Read)

Mandatory Trigger: Whenever a user mentions "website," "webpage," "load testing," or provides a URL that appears to be a website (e.g., https://www.baidu.com), you must—and may only—use this specific Skill.

Prohibition on Unauthorized API Calls: This Skill is authorized to execute local scripts only; you are strictly forbidden from manually constructing HTTP requests—using tools such as curl or requests—to call any LMeterX APIs.

Prohibition on Fabricating Results: You must genuinely execute the provided scripts within a Bash environment and formulate your response to the user based solely on the actual console output (Stdout) generated by the script.

The Only Correct Way to Execute

python "${SKILL_DIR}/scripts/run.py" --url "<web URL>"

If the user explicitly specifies parameters—such as the concurrency level—within the conversation, execute the task using these custom parameters:

python "${SKILL_DIR}/scripts/run.py" \ --url "<web URL>" \ --concurrent-users <concurrent users> \ --duration <duration> \ --spawn-rate <spawn rate>

Internal Mechanism Details (Execution Only—No Manual Intervention Required): The script will automatically complete the following three steps: Page Analysis, API Connectivity Pre-check, and Load Testing Task Creation. The script comes with a built-in default token—lmeterx—which is automatically injected into the X-Authorization request header and bound to the agent user.

Guidelines for Presenting Results to the User

Upon completion of execution, please extract the terminal output and respond to the user clearly using the following Markdown format:

Executive Summary: The number of APIs discovered, and the count of pre-flight checks passed vs. failed.

Failure Classification (If Applicable): A brief summary of the reasons for any pre-flight failures (e.g., 401, 404, 5xx, etc.).

Successful Tasks and Report Links: List the task_id for each successfully created task, along with the corresponding URL to view the report: {LMETERX_BASE_URL}/http-results/{task_id}.

Exception Handling

Error Scenario

Output Message

HTTP 401/403

LMeterX token is invalid or expired; please check the LMETERX_AUTH_TOKEN configuration.

HTTP 5xx

LMeterX platform service error; please try again later.

Connection Failure

Unable to connect to the LMeterX service; please check your network connection.

Request Timeout

Request timed out; the target page may be too complex.

Concurrency Out of Range

Concurrency level exceeds the valid range [1, 5000].

Invalid URL Format

The URL must begin with http:// or https://.