AI agents call build_agent_context_pack to retrieve information from Todos without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool constructs a deterministic context pack for AI agents and returns it as JSON or Markdown. It explicitly states it never calls hosted services, indicating no side effects. This is a read/query operation that assembles and returns information.
From the tool's definition 'Returns JSON or Markdown and never calls hosted services' — the tool builds and returns a context pack without modifying any data or executing external operations.
Attacks that exploit this kind of access
Build a deterministic local agent context pack for Codex, Claude Code, Takumi, or generic agents. Returns JSON or Markdown and never calls hosted services. It is categorised as a Read tool in the Todos MCP Server, which means it retrieves data without modifying state.
Register the Todos MCP server in PolicyLayer and add a rule for build_agent_context_pack: 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 Todos. Nothing to install.
build_agent_context_pack 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 build_agent_context_pack 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 build_agent_context_pack. 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.
build_agent_context_pack is provided by the Todos MCP server (@hasna/todos). 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 →