Persist a structured compact-checkpoint journal entry just before the host runs auto-summarization on a long context window. Scans the prefix (up to 8000 chars) and extracts: open_threads (questions without clean answers), last_user_intent, decisions_in_flight, tool_failures. Writes one agent_jou...
AI agents use compact_checkpoint to create or update resources in Celiums Memory — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Celiums Memory environment.
The tool creates and writes a new journal entry to persistent storage (agent_journal table). This is a Write operation—it creates/modifies data reversibly. Severity is medium because the checkpoint data itself is benign metadata about agent state (open threads, intents, decisions), not sensitive or critical business data, though inappropriate writes could clutter the journal or overwrite important context.
From the tool's definition Persist a structured compact-checkpoint journal entry...Writes one agent_journal row tagged
Documented attack patterns abuse exactly the kind of access compact_checkpoint gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Celiums Memory, and nothing reaches the server without passing your rules. This is the rule we recommend for compact_checkpoint:
{
"version": "1",
"default": "deny",
"tools": {
"compact_checkpoint": {
"limits": [
{
"counter": "compact_checkpoint_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
}compact_checkpoint 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.
Persist a structured compact-checkpoint journal entry just before the host runs auto-summarization on a long context window. Scans the prefix (up to 8000 chars) and extracts: open_threads (questions without clean answers), last_user_intent, decisions_in_flight, tool_failures. Writes one agent_journal row tagged. It is categorised as a Write tool in the Celiums Memory MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Celiums Memory MCP server in PolicyLayer and add a rule for compact_checkpoint: 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 Celiums Memory. Nothing to install.
compact_checkpoint 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 compact_checkpoint 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 compact_checkpoint. 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.
compact_checkpoint is provided by the Celiums Memory MCP server (terrizoaguimor/celiums-memory). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Celiums Memory, 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.
62 Celiums Memory tools catalogued and risk-classified — across an index of 43,000+ MCP servers.