Scan ~/.claude/projects for projects Claude Code has touched on this machine, decode each directory name back to its absolute path, and report which ones still exist plus session-file count and last activity. With add_as_subprojects=true, every existing project is registered as a subproject in on...
AI agents call discover_claude_sessions to retrieve information from Trace without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
Primary function is information retrieval—discovering and reporting on existing Claude projects and their metadata (path, session count, last activity). While the optional add_as_subprojects flag can register subprojects, this remains a Write-like metadata registration within the tool's own state management, not a destructive or risky external operation.
From the tool's definition Tool explicitly states it 'Reads local filesystem' and 'Scans ~/.claude/projects' to report on projects and sessions. The add_as_subprojects mutation is secondary and operates within the MCP topology store (creating references, not modifying external state).
Risk signalsBulk/mass operation — affects multiple targets
Documented attack patterns abuse exactly the kind of access discover_claude_sessions gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Trace, and nothing reaches the server without passing your rules. This is the rule we recommend for discover_claude_sessions:
{
"version": "1",
"default": "deny",
"tools": {
"discover_claude_sessions": {}
}
}discover_claude_sessions is read-only, so it stays allowed — but everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Scan ~/.claude/projects for projects Claude Code has touched on this machine, decode each directory name back to its absolute path, and report which ones still exist plus session-file count and last activity. With add_as_subprojects=true, every existing project is registered as a subproject in one call — useful for spinning up multi-repo intelligence after a fresh clone. Reads local filesystem; with add_as_subprojects=true also mutates topology store. Returns JSON: { projects: [{ path, sessions, lastActivity }], total }. It is categorised as a Read tool in the Trace MCP Server, which means it retrieves data without modifying state.
Register the Trace MCP server in PolicyLayer and add a rule for discover_claude_sessions: 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 Trace. Nothing to install.
discover_claude_sessions 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 discover_claude_sessions 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 discover_claude_sessions. 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.
discover_claude_sessions is provided by the Trace MCP server (nikolai-vysotskyi/trace-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Deterministic rules across all 178 Trace tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.
Free to start. No card required.
178 Trace tools catalogued and risk-classified — across an index of 42,500+ MCP servers.