Track NPC relationships, memories, and social interactions. 🎭 SOCIAL AI WORKFLOW: 1. get_context - Before NPC dialogue, get relationship + memory summary 2. Inject into system prompt for informed roleplay 3. record_memory - After significant interactions 4. update_relationship - When familiarity...
AI agents use npc_manage to create or update resources in Rpg — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Rpg environment.
npc_manage creates and modifies NPC data (memories, relationships, disposition states) persistently in the SQLite backend. While it affects game mechanics and NPC behavior, it does not execute arbitrary code, delete data irreversibly, or move financial resources.
From the tool's definition The tool modifies game state through 'record_memory' and 'update_relationship' operations that change NPC relationship progression (familiarity/disposition levels) and memory records. These are reversible write operations within the game system.
Documented attack patterns abuse exactly the kind of access npc_manage gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Rpg, and nothing reaches the server without passing your rules. This is the rule we recommend for npc_manage:
{
"version": "1",
"default": "deny",
"tools": {
"npc_manage": {
"limits": [
{
"counter": "npc_manage_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
}npc_manage 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.
Track NPC relationships, memories, and social interactions. 🎭 SOCIAL AI WORKFLOW: 1. get_context - Before NPC dialogue, get relationship + memory summary 2. Inject into system prompt for informed roleplay 3. record_memory - After significant interactions 4. update_relationship - When familiarity/disposition changes 📊 RELATIONSHIP PROGRESSION: Familiarity: stranger → acquaintance → friend → close_friend (or rival/enemy) Disposition: hostile → unfriendly → neutral → friendly → helpful 🗣️ SPATIAL INTERACTIONS: - interact: Volume affects who hears (WHISPER/TALK/SHOUT) - Eavesdroppers roll Stealth vs Perception - Memories auto-recorded for participants 💡 AI TIP: Always call get_context before generating NPC dialogue! Response includes formatted summary for prompt injection. Actions: create, get_full_context, get_relationship, update_relationship, record_memory, get_history, get_recent, get_context, interact. It is categorised as a Write tool in the Rpg MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Rpg MCP server in PolicyLayer and add a rule for npc_manage: 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 Rpg. Nothing to install.
npc_manage 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 npc_manage 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 npc_manage. 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.
npc_manage is provided by the Rpg MCP server (mnehmos/mnehmos.rpg.mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Rpg, 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.
47 Rpg tools catalogued and risk-classified — across an index of 43,000+ MCP servers.