policy_decisions

Audit and understand recent policy evaluation decisions (allow/deny/escalate) for workspace oversight and compliance verification.

Instructions

View recent policy evaluation decisions (allow/deny/escalate) for this workspace. Useful for auditing and understanding what the policy engine has been doing.

Input Schema

TableJSON Schema

Name	Required	Description	Default
`limit`	No	Max decisions to return (default 10)
`decision`	No	Filter by decision type

Implementation Reference

openterms_mcp_server.py:345-365 (handler)

The handler implementation for the 'policy_decisions' tool, which fetches recent policy decisions from the /v1/policy/decisions endpoint.

elif name == "policy_decisions":
    params = {"limit": min(arguments.get("limit", 10), 50)}
    if arguments.get("decision"):
        params["decision"] = arguments["decision"]
    resp = client.get("/v1/policy/decisions", params=params, headers=_headers())
    if resp.status_code == 200:
        data = resp.json()
        decisions = data.get("decisions", [])
        if not decisions:
            return "No policy decisions recorded yet."
        lines = [f"Recent policy decisions ({len(decisions)}):"]
        for d in decisions:
            icon = {"allow": "✅", "deny": "🚫", "escalate": "⏸️"}.get(d.get("decision"), "❓")
            reasons = d.get("reasons", [])
            reason_str = f" — {reasons[0]}" if reasons else ""
            lines.append(
                f"  {icon} {d.get('decision', '?').upper()} v{d.get('profile_version', '?')} "
                f"receipt={d.get('receipt_id', 'n/a')[:12]}...{reason_str} ({d.get('evaluated_at', '?')})"
            )
        return "\n".join(lines)
    return _format_error(resp)

openterms_mcp_server.py:106-118 (schema)

The MCP tool schema definition for 'policy_decisions', including the input schema for limit and decision filter.

    "name": "policy_decisions",
    "description": (
        "View recent policy evaluation decisions (allow/deny/escalate) for this workspace. "
        "Useful for auditing and understanding what the policy engine has been doing."
    ),
    "inputSchema": {
        "type": "object",
        "properties": {
            "limit": {"type": "integer", "description": "Max decisions to return (default 10)"},
            "decision": {"type": "string", "enum": ["allow", "deny", "escalate"], "description": "Filter by decision type"},
        },
    },
},

Openterms-mcp