Security audit
Security checks for vulnerabilities and agentic risk
This appears to be a legitimate workflow generator, but it ships templates and examples that can create agents with broad shell, file-write, network, and persistent-memory powers without enough built-in safeguards.
Install only if you intentionally want a skill that generates runnable AWP workflow projects. Use it in a dedicated workspace, review the generated plan before approving file creation, and carefully inspect any generated YAML, prompts, MCP tools, Cloudflare adapter, memory settings, and secrets configuration before running. Be especially cautious with shell.execute, infra.run_command, http.request, file.write, broadcast messaging, persistent memory, and any workflow that sends prompts or outputs to third-party LLM/API providers.
"error": f"Blocked: command matches dangerous pattern '{pattern}'",
}
try:
result = subprocess.run(
command, shell=True, capture_output=True, text=True, timeout=timeout,
)
return {"error": f"Blocked: command matches dangerous pattern '{pattern}'",
}
try:
result = subprocess.run(
command, shell=True, capture_output=True, text=True, timeout=timeout,
)
return {Detected: suspicious.prompt_injection_instructions