AI agents call recent_changes to retrieve information from Srclight without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
Despite the empty description, the name 'recent_changes' in the context of a code indexing server almost certainly retrieves historical information about code modifications from version control (git). This is a Read operation—it queries data without modifying anything. The server is explicitly designed for code navigation and history tracking, and all sibling tools appear to be read-only searches and analysis.
From the tool's definition Tool name 'recent_changes' combined with server description emphasizing 'track git history' and the context of sibling tools like 'blame_symbol', 'changes_to', and 'detect_changes' which are all query/navigation operations.
Documented attack patterns abuse exactly the kind of access recent_changes gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Srclight, and nothing reaches the server without passing your rules. This is the rule we recommend for recent_changes:
{
"version": "1",
"default": "deny",
"tools": {
"recent_changes": {}
}
}recent_changes 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.
recent_changes. It is categorised as a Read tool in the Srclight MCP Server, which means it retrieves data without modifying state.
Register the Srclight MCP server in PolicyLayer and add a rule for recent_changes: 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 Srclight. Nothing to install.
recent_changes 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 recent_changes 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 recent_changes. 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.
recent_changes is provided by the Srclight MCP server (srclight/srclight). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Deterministic rules across all 42 Srclight tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.
Free to start. No card required.
42 Srclight tools catalogued and risk-classified — across an index of 42,500+ MCP servers.