TraceGuardDocs

Authentication & tokens

The four token types and how scopes work.

TraceGuard issues bearer tokens with a typed prefix so you can tell at a glance what a token is for:

TypePrefixUse it for
Personal access tokentgp_Acting as yourself from scripts or the CLI.
Service account tokentgs_Automation owned by a service account.
Deploy tokentgd_CI uploads, scoped to a single project.
MCP tokentgm_Connecting an MCP client (Claude, an IDE).

Send a token as a bearer header:

curl -H "Authorization: Bearer tgd_xxxxxxxx" \
  "https://app.traceguard.io/api/v1/organizations/$ORG/projects/$PROJECT/sboms"

Scopes

A token's authority is a set of scopes — the same resource:action permission IDs used by roles. Scopes support wildcards:

  • * — every permission the granter has.
  • sboms:* — every action on a resource.
  • *:read — a verb across resources.

Scopes are always intersected with the authority of whoever created the token, and enforced again at the database boundary. A token can never do more than its creator could — and never read or write across organization boundaries.

Tokens are shown once at creation. Store them in a secrets manager, scope them to the minimum they need, and rotate them if exposed. Never commit a token to source control.

On this page