- SECURITY.md: spell out that agent.key encryption is off-host/tamper protection, not confidentiality against local root; 0600 perms are the real control (L2). - agent_registry.py: reword the per-message session-token "TODO" to an accepted limitation (defence-in-depth, not a bug) (L7). - AGENT_AUDIT.md: record H1/H2/H3 and L1-L7 status, with L4 (cap OR-merge), L5 (single-worker), L6 (windows .bat) noted as accepted by-design/documented. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2.8 KiB
Security Policy
Supported Versions
ServerKit is under active development. Security fixes are applied to the latest
release line and the main branch.
| Version | Supported |
|---|---|
| 1.6.x | ✅ |
| < 1.6 | ❌ (please upgrade) |
The agent is versioned independently (agent-vX.Y.Z); always run a recent
agent build, as several Windows service and credential-handling fixes landed in
the 1.6.x line.
Reporting a Vulnerability
Please report security issues privately — do not open a public issue for anything exploitable.
- Preferred: open a GitHub Security Advisory
- We aim to acknowledge reports within a few days and to provide a remediation timeline after triage.
Please include affected version(s), reproduction steps, and impact. Coordinated disclosure is appreciated — give us a reasonable window to ship a fix before any public write-up.
Agent Trust Model
The multi-server agent is powerful by design — operators should understand its trust boundaries:
agent.keyis a host-equivalent secret. Agent API credentials are stored AES-256-GCM encrypted under a key derived from host-stable identifiers (hostname + machine ID on Linux, hostname + computer name on Windows). Because that key is derived only from values available on the host itself, the encryption is at-rest tamper-resistance / off-host-exfil protection (e.g. a leaked backup) — not confidentiality against a local root/SYSTEM user, who can re-derive the key. Anyone who can read this file on the host can recover the credentials. The0600file permissions are the real access control; protect it like a root/SYSTEM secret.- Remote command execution is gated. Arbitrary command execution
(
system:exec) and interactive PTY sessions are controlled by the agent'sFeatures.Execflag, which is off by default. Enable it only on servers where you intend the panel to run shell commands. - Transport & connection controls. Agents authenticate to the panel with
per-connection HMAC-SHA256 (with nonce/replay protection and a timestamp-skew
check), and the panel enforces a per-server IP allowlist. Use
wss://(TLS-terminated) in production. SERVERKIT_INSECURE_TLS=truedisables certificate verification for all agent connections. It is intended for local development/testing only — never set it in production.
For a detailed internal audit of the panel, see SECURITY_AUDIT.md.
Deployment Note
The agent gateway keeps all connected-agent state in-memory in a single process. Run the panel with a single gevent-websocket worker; multi-worker deployments can misroute agent commands. See docs/ARCHITECTURE.md.