Retrieve traces for a specific service, with an option to filter for traces that contain errors. Returns a list of traces, each containing: traceID (unique trace identifier), latency_ms (total trace duration in milliseconds), has_error (boolean indicating if the trace contains errors), and sequen...
AI agents call get_traces to retrieve information from K8s Observability MCP without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool only queries and returns observability data (traces, latency, error flags, service call sequences) without modifying state, executing code, or triggering external operations. It is a read-only data retrieval function typical of monitoring/observability systems. Low severity because trace data is generally non-sensitive metadata about service interactions, and retrieval alone poses minimal risk of misuse.
From the tool's definition Tool retrieves traces for a specific service with optional error filtering. Description explicitly states it 'Retrieve traces' and 'Returns a list of traces' with no modification, deletion, or execution capabilities mentioned.
Attacks that exploit this kind of access
Retrieve traces for a specific service, with an option to filter for traces that contain errors. Returns a list of traces, each containing: traceID (unique trace identifier), latency_ms (total trace duration in milliseconds), has_error (boolean indicating if the trace contains errors), and sequence (string showing the service call chain, e.g.,. It is categorised as a Read tool in the K8s Observability MCP MCP Server, which means it retrieves data without modifying state.
Register the K8s Observability MCP server in PolicyLayer and add a rule for get_traces: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches K8s Observability MCP. Nothing to install.
get_traces is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the get_traces rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for get_traces. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
get_traces is provided by the K8s Observability MCP server (martinimarcello00/k8s-observability-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →