Apus supports additive SecretRefs so supported credentials do not need to live as plaintext in configuration.

Plaintext still works. SecretRefs are opt-in per credential.

Plaintext credentials remain agent-readable when they sit in files the agent can inspect, including apus.json, .env, retired auth-profile JSON archives, or generated agents/*/agent/models.json files. SecretRefs reduce that local blast radius once every supported credential is migrated and apus secrets audit --check reports no plaintext residue.

This page is an index. Secrets management is documented on five pages, one per reader job. Open the page that matches your task.

Secrets pages

Page Read it when
Secrets runtime model Owner isolation, sentinel injection, the agent-access boundary, and active-surface filtering.
SecretRef contract and provider config The SecretRef contract, id grammars, validation rules, and the env, file, exec, and store provider blocks.
Shared secret store and egress proxy The shared secret store, the secret egress proxy and its traffic allowlist, and file-backed API keys.
Secrets integration examples Exec provider recipes for 1Password, Bitwarden, Vault, pass, and sops, plus MCP and sandbox SSH.
Secrets operations and behavior Supported surfaces, precedence, activation triggers, degraded signals, and the audit and configure workflow.

Where each section moved

Every section, tab, step, and accordion title from the previous single-page version keeps its anchor here, so an existing link such as /gateway/secrets#shared-secret-store still resolves. Each entry points at the page that now holds the content.

Related