Aletheia Runtime Safety & Scope Creep Filter runs on your own machine — the client starts it, so there is no endpoint to ping. 65 installs a week from npm. Last commit 12 Sep 2026.
Deterministic pre-execution filter that blocks known scope-creep and prompt-injection tool calls
We read the source, 4 d ago · rules 995398eba69f
Things with no honest explanation: a promise that contradicts the code, code that runs at install time while hiding what it does, data leaving the machine.
- **Filesystem Excursions**: A request to "inspect git diff" escalating to deleting untracked files with \`git clean -fdx\` or \`rm -rf\`.
What this server is able to do. For an MCP server this is often the job itself — a terminal server runs commands because that is what it is for. Listed so you know what you are plugging in, not as an accusation.
const hasDelete = /\bDELETE\s+FROM\b/i.test(repSpace) || /\bDELETE\s+FROM\b/i.test(repCollapsed);
this.childProcess = spawn(downstreamCommand, downstreamArgs, {
trimmed.includes("/.ssh/") ||
Is this your server and something here is wrong? Tell us — corrections are free and do not require a plan.
Code changes quietly between releases, and nobody reads the diff of a dependency. We do, on every release — watch Aletheia Runtime Safety & Scope Creep Filter and you get told the day something new turns up.
This server runs on your own machine — install it with the package manager and the client starts it for you. Package name taken from the official registry entry.
claude mcp add aletheia-mcp -- npx -y aletheia-mcp
{
"mcpServers": {
"aletheia-mcp": {
"command": "npx",
"args": [
"-y",
"aletheia-mcp"
]
}
}
}
[mcp_servers.aletheia-mcp]
command = "npx"
args = ["-y", "aletheia-mcp"]
{
"mcpServers": {
"aletheia-mcp": {
"command": "npx",
"args": [
"-y",
"aletheia-mcp"
]
}
}
}
{
"servers": {
"aletheia-mcp": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"aletheia-mcp"
]
}
}
}
Optional settings from the registry entry, left out of the block above:
ALETHEIA_ALLOW_WRITE (Set to "true" to authorize filesystem writes (equivalent to --allow-write)), ALETHEIA_ALLOW_NETWORK (Set to "true" to authorize outbound network requests (equivalent to --allow-network)), ALETHEIA_ALLOW_LOOPBACK (Set to "true" to authorize requests to localhost / 127.0.0.1 (equivalent to --allow-loopback)), ALETHEIA_OPERATOR_SECRET (Operator token required to loosen an already-running session's mandate at runtime).
Security gateway for MCP agents: blocks prompt-injection-driven tool calls before they execute.
Governed MCP: agent audit, provenance, deterministic checks, and receipt-backed FragGate execution.
Deterministic runtime safety for AI agents: scan PII, gate tool actions, verify LLM output.
Deterministic pre-execution safety certification for autonomous AI agents.
Evaluate RAG retrieval, agent tool calls and prompt regressions with deterministic metrics.
Prompt injection + PII detection for LLM apps and agent pipelines. Deterministic, sub-25ms.
MCP security guard + package manager: trust-scored installs, blocks prompt injection and rug-pulls.
Deterministic pre-execution verification tools for AI agents with x402 USDC pay-per-call on Base
Answers built from our own checks of this server.