Show the call stack (backtrace).
AI agents call backtrace to retrieve information from gdb and rr Debugging without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This is a read-only debugging operation that retrieves and displays the current call stack during program debugging. It does not modify state, execute arbitrary code, or cause destructive changes. While the broader GDB server permits Execute-level operations (like 'exec_command', 'run', 'start_replay_session'), the backtrace tool itself is confined to passive inspection.
From the tool's definition The tool 'backtrace' returns the call stack and has no side effects. The description states it 'Show[s] the call stack' — purely an inspection/retrieval operation. No modification, deletion, or execution of external code occurs.
Attacks that exploit this kind of access
Show the call stack (backtrace). It is categorised as a Read tool in the gdb and rr Debugging MCP Server, which means it retrieves data without modifying state.
Register the gdb and rr Debugging MCP server in PolicyLayer and add a rule for backtrace: 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 gdb and rr Debugging. Nothing to install.
backtrace 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 backtrace 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 backtrace. 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.
backtrace is provided by the gdb and rr Debugging MCP server (schuay/gdb-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 →