AI agents call feed_digest to retrieve information from Devfeed without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The feed_digest tool performs pure data retrieval: it queries multiple developer community platforms (Hacker News, Lobsters, Reddit, Dev.to) and returns filtered content to the user. No data is created, modified, deleted, or executed. This is a straightforward read operation similar to searching or fetching content.
From the tool's definition Tool provides a 'personalized digest' that 'filtered by your interests' — it retrieves and aggregates content from multiple sources without modifying or deleting data. The description explicitly indicates retrieval of news feeds with no side effects.
Attacks that exploit this kind of access
Your personalized digest across Hacker News, Lobsters, Reddit, and Dev.to — filtered by your interests from ~/.devfeed.yaml. It is categorised as a Read tool in the Devfeed MCP Server, which means it retrieves data without modifying state.
Register the Devfeed MCP server in PolicyLayer and add a rule for feed_digest: 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 Devfeed. Nothing to install.
feed_digest 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 feed_digest 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 feed_digest. 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.
feed_digest is provided by the Devfeed MCP server (upgpt-ai/devfeed-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 →