Save current state to disk for crash recovery and session resume. Checkpoints include all fragments, dedup index, co-access patterns, and custom metadata. Stored as gzipped JSON (~50-200 KB). Args: task_description: What the agent is working on current_step: Where in the task it currently is
AI agents use checkpoint_state to create or update resources in Entroly Context Engine — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Entroly Context Engine environment.
This tool creates or modifies data on disk (checkpoint files) but does so reversibly—checkpoints can be overwritten or deleted without permanent consequence. It does not delete data irreversibly, execute arbitrary code, or move money. The blast radius is limited to the integrity of checkpoint recovery files.
From the tool's definition Tool description states 'Save current state to disk' and 'Checkpoints include all fragments, dedup index, co-access patterns, and custom metadata. Stored as gzipped JSON'.
Documented attack patterns abuse exactly the kind of access checkpoint_state gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Entroly Context Engine, and nothing reaches the server without passing your rules. This is the rule we recommend for checkpoint_state:
{
"version": "1",
"default": "deny",
"tools": {
"checkpoint_state": {
"limits": [
{
"counter": "checkpoint_state_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
}checkpoint_state stays usable, but capped — an agent stuck in a loop can't make hundreds of changes a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Save current state to disk for crash recovery and session resume. Checkpoints include all fragments, dedup index, co-access patterns, and custom metadata. Stored as gzipped JSON (~50-200 KB). Args: task_description: What the agent is working on current_step: Where in the task it currently is. It is categorised as a Write tool in the Entroly Context Engine MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Entroly Context Engine MCP server in PolicyLayer and add a rule for checkpoint_state: 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 Entroly Context Engine. Nothing to install.
checkpoint_state is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the checkpoint_state 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 checkpoint_state. 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.
checkpoint_state is provided by the Entroly Context Engine MCP server (juyterman1000/entroly). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Entroly Context Engine, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
52 Entroly Context Engine tools catalogued and risk-classified — across an index of 43,000+ MCP servers.