Predict which files are most likely to contain bugs. Multi-signal scoring: git churn, fix-commit ratio, complexity, coupling, PageRank importance, author count. Each prediction includes a numeric score, risk bucket (low/medium/high/critical) AND a confidence_level (low/medium/high/multi_signal) c...
AI agents call predict_bugs 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.
This tool performs static analysis and machine learning-based prediction over existing code and git history. It retrieves and analyzes data to produce insights but does not execute code, modify files, delete data, or trigger external operations. The 'Read-only' declaration and return-only behavior (predictions with scoring) confirm classification as Read.
From the tool's definition Tool description explicitly states 'Read-only' and describes analysis-only operations: 'Predict which files are most likely to contain bugs' via 'git churn, fix-commit ratio, complexity, coupling, PageRank importance, author count'.
Documented attack patterns abuse exactly the kind of access predict_bugs 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 predict_bugs:
{
"version": "1",
"default": "deny",
"tools": {
"predict_bugs": {}
}
}predict_bugs 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.
Predict which files are most likely to contain bugs. Multi-signal scoring: git churn, fix-commit ratio, complexity, coupling, PageRank importance, author count. Each prediction includes a numeric score, risk bucket (low/medium/high/critical) AND a confidence_level (low/medium/high/multi_signal) counting how many independent signals actually fired. Result envelope includes _methodology disclosure. Cached for 1 hour; use refresh=true to recompute. Requires git. Use for proactive bug hunting. For complexity+churn hotspots only use get_risk_hotspots instead. Read-only. Returns JSON: { predictions: [{ file, score, risk, confidence_level, signals }], 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 predict_bugs: 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.
predict_bugs 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 predict_bugs 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 predict_bugs. 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.
predict_bugs 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.