# one.secureai/redaction

> Takes the people out of text before a model sees it, and puts them back in the reply.

- name: one.secureai/redaction
- version: 1.0.0
- connection class: R2 (oauth_consent_once)
- trust: 93/100
- quality: 94/100 (strong)
- owner: not claimed
- descriptor: https://api.protogrid.dev/v1/servers/one.secureai%2Fredaction
- tools: https://api.protogrid.dev/v1/servers/one.secureai%2Fredaction/tools

**An agent can connect after a one-time OAuth consent by a human.** Remote endpoint behind OAuth 2. Tokens stay with the agent; the registry only exposes the authorization metadata.

## Connection (mcpServers)

```json
{
  "mcpServers": {
    "redaction": {
      "type": "http",
      "url": "https://secureai.one/mcp",
      "headers": {
        "Authorization": "Bearer ${REDACTION_TOKEN}"
      }
    }
  }
}
```

Placeholders to fill: `${REDACTION_TOKEN}`.

## Trust

- hygiene: 100
- liveness: 100
- freshness: 85
- provenance: 85
- drivers: +repository +dns-namespace +website +reachable +uptime-99% ~updated-32d-ago
- Derived from observable signals (official registry feed and our own credential-free probes); no code audit performed.

## Quality 94/100 (strong)

### protocol: n/a (weight 25)

- n/a `protocol.modern`, `protocol.stateless`, `protocol.transport`, `protocol.list_ttl`: no remote answered a protocol handshake (not probed yet, unreachable, or local-only)

### authorization: 83 (weight 15)

- pass `auth.prm`: Protected Resource Metadata (RFC 9728) published
- pass `auth.as_metadata`: authorization server metadata reachable
- warn `auth.cimd`: only Dynamic Client Registration, deprecated in 2026-07-28 in favour of CIMD
- n/a `auth.secret_in_url`: no templated URL

### tool hygiene: 100 (weight 30)

- pass `tools.descriptions`: every tool has a description
- pass `tools.description_length`: descriptions are concise
- pass `tools.schemas`: every tool has an object input schema
- pass `tools.annotations`: 2 of 2 tools declare readOnlyHint or destructiveHint
- pass `tools.directory_hints`: every tool declares readOnlyHint, destructiveHint, idempotentHint and openWorldHint
- pass `tools.token_cost`: about 352 tokens to load every tool
- pass `tools.api_dump`: tools are not a one-to-one API dump

### stability: n/a (weight 15)

- n/a `stability.changes`, `stability.rug_pull`: tool history not recorded yet

### dependencies: n/a (weight 15)

- n/a `deps.known_vulns`, `deps.mcp_sdk_version`, `deps.resolvable`: no npm or PyPI package

- tools: 2, about 352 tokens to load them all
- badge: [![protogrid quality](https://protogrid.dev/badge/one.secureai/redaction.svg)](https://protogrid.dev/servers/one.secureai/redaction)
- Checks run on what our credential-free, read-only probes observe; tools are never called and no code audit is performed.

### Tool changes

_No tool definition changes recorded._

## Remotes

- https://secureai.one/mcp (streamable-http, auth oauth2, reachable true, uptime30d 0.9915254)

## Packages

_none_

## Tools (2)

- `redact`: Replace the people in a piece of text — names, emails, phone numbers, addresses, card and account numbers — with plausible stand-ins, before the text is sent to a model or a third party. Returns the redacted text and a map. Keep the map: it is the only way to put the real values back, and it is not stored anywhere.
- `restore`: Reverse a redaction using the map it returned. Use this on a model's reply so the reader sees what they actually wrote rather than the stand-ins.