The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to generate load test scripts based on user-provided parameters, which could be exploited to inject malicious code into the generated scripts if the input is not properly sanitized.
Ingestion points: User requests defining target endpoints, concurrent user counts, and test scenario parameters (e.g., Example 1 in SKILL.md and descriptions in scripts/README.md).
Boundary markers: Absent. The instructions do not define delimiters or specific warnings to ignore instructions embedded in the user-provided parameters.
Capability inventory: The skill has access to Write for file creation and Bash (restricted to specific prefixes) for execution.
Sanitization: Absent. There are no explicit instructions or mechanisms mentioned for validating or escaping user input before interpolation into script templates.
[DYNAMIC_EXECUTION]: The skill generates executable scripts (k6 JavaScript, JMeter JMX, etc.) at runtime based on templates and user input, which are then processed by an executor script.
Evidence: scripts/README.md lists k6_script_generator.py and load_test_executor.py. SKILL.md instructs the agent to "Generate appropriate load test scripts" and "Execute load tests and capture metrics."
[COMMAND_EXECUTION]: The skill executes shell commands using the Bash tool to run third-party load testing utilities.
Evidence: The allowed-tools section in SKILL.md specifies Bash(k6:*), Bash(jmeter:*), Bash(artillery:*), and Bash(performance:*) to allow the agent to run the necessary testing tools.

running-load-tests