api-load-tester
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill generates and executes shell commands for performance testing tools (k6, wrk, autocannon). While this is the primary purpose of the skill, it involves high-volume network traffic generation which could be misused if directed at unauthorized targets.
- Evidence: `k6 run --vus 100 --duration 60s
- <<'EOF'...
in Example 1, and instructions to executewrkandautocannon` commands. - [INDIRECT_PROMPT_INJECTION]: The skill has a significant attack surface for indirect prompt injection because it is designed to ingest and process untrusted external data (OpenAPI specs, route files, and source code files) to generate test logic.
- Ingestion points: Reads OpenAPI/Swagger specs and source code (TypeScript, Python, Go) to generate payloads and test flows.
- Boundary markers: None specified in the instructions to prevent the agent from following instructions embedded within the API documentation or code comments.
- Capability inventory: Execution of network load testing tools (k6, wrk) and command-line operations.
- Sanitization: No explicit sanitization or validation of the ingested API definitions is mentioned before they are used to generate executable test scripts.
- [REMOTE_CODE_EXECUTION]: The skill generates scripts (JavaScript for k6 and autocannon) at runtime based on external input. While these are executed within the context of the load testing tool, it represents dynamic script generation and execution.
Audit Metadata