AI agents use score_memories to create or update resources in Roampal Core — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Roampal Core environment.
Scoring memories implies updating/annotating existing memory entries with outcome scores. This is a reversible data modification (Write), not a read-only operation. Severity is medium because misuse could corrupt the outcome-based memory system that guides AI coding tool behavior, but it doesn't delete data outright.
From the tool's definition 'Score previous exchange outcomes' — scoring modifies stored memory records with outcome evaluations
Documented attack patterns abuse exactly the kind of access score_memories gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Roampal Core, and nothing reaches the server without passing your rules. This is the rule we recommend for score_memories:
{
"version": "1",
"default": "deny",
"tools": {
"score_memories": {
"limits": [
{
"counter": "score_memories_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
}score_memories 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.
Score previous exchange outcomes. It is categorised as a Write tool in the Roampal Core MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Roampal Core MCP server in PolicyLayer and add a rule for score_memories: 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 Roampal Core. Nothing to install.
score_memories 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 score_memories 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 score_memories. 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.
score_memories is provided by the Roampal Core MCP server (roampal-ai/roampal-core). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Deterministic rules across all 6 Roampal Core tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.
Free to start. No card required.
6 Roampal Core tools catalogued and risk-classified — across an index of 42,500+ MCP servers.