← Back to search

Agent Guardrail MCP Server

eren-solutions Scanned 5d ago

Action-level governance for AI agents -- control what they DO, not what they SAY

MCP Registry ↗
C
73.5 / 100

Versions

0.1.2latest
first seen May 19, 2026
PermissionsTool SafetyAuthAnnotationsCode QualityStabilitySpecVuln HistoryAuthorTransparencyCommunity

Tools 6

evaluate_action
annotations: none low

Evaluate whether an agent action is allowed by the guardrail policy. Call this BEFORE executing any tool, shell command, or HTTP request.

target string agent_id string cost_usd string tool_name string action_type string
register_agent
annotations: none low

Register a new agent with the guardrail system. Must be called before evaluate_action.

name string framework string description string
list_agents
annotations: none low

List all agents currently registered in the guardrail system. Shows active and killed agents. Returns: Array of agent records with fields: id, name, framework, description, status, created_at.

get_stats
annotations: none low

Retrieve aggregate statistics from the guardrail system including evaluation counts, policy data, and agent metrics. Returns: Statistics object with fields: total_evaluations, allowed_count, denied_count, agent_count, policy_count.

kill_agent
annotations: none low

Emergency kill-switch: immediately block ALL actions for an agent. Use when an agent is misbehaving or compromised.

agent_id string
unkill_agent
annotations: none low

Re-enable a previously killed agent, restoring its ability to have actions evaluated.

agent_id string

Permissions 5

network medium
Server uses network capabilities via: urllib
filesystem low
Server uses filesystem capabilities via: open(), os, pathlib, tempfile
shell high
Server uses shell capabilities via: subprocess
database medium
Server uses database capabilities via: sqlite3
env_vars low
Server uses env_vars capabilities via: os.environ

Scan Findings 25

low
Tool 'evaluate_action' has no annotations annotation_checker · 100%
low
Tool 'register_agent' has no annotations annotation_checker · 100%
low
Tool 'list_agents' has no annotations annotation_checker · 100%
low
Tool 'get_stats' has no annotations annotation_checker · 100%
low
Tool 'kill_agent' has no annotations annotation_checker · 100%
low
Tool 'unkill_agent' has no annotations annotation_checker · 100%
info
Sandbox failed to start for behavioral verification behavioral_verifier · 100%
info
pyproject.toml metadata manifest_parser · 100%
info
Tool: evaluate_action manifest_parser · 90%
info
Tool: register_agent manifest_parser · 90%
info
Tool: list_agents manifest_parser · 90%
info
Tool: get_stats manifest_parser · 90%
info
Tool: kill_agent manifest_parser · 90%
info
Tool: unkill_agent manifest_parser · 90%
info
Transport: streamable-http manifest_parser · 80%
info
Required env vars (14) manifest_parser · 80%
info
Sandbox failed to start for output poisoning scan output_poisoning · 100%
medium
Permission: network access detected permission_analyzer · 80%
low
Permission: filesystem access detected permission_analyzer · 80%
high
Permission: shell access detected permission_analyzer · 95%
medium
Permission: database access detected permission_analyzer · 90%
low
Permission: env_vars access detected permission_analyzer · 90%
critical
Tool poisoning in 'evaluate_action': Cross-tool sequencing directive poisoning · 85%
info
No dependency files found for SBOM generation sbom_generator · 100%
medium
No build provenance detected (SLSA L0) slsa_assessor · 90%