vultr/openclaw
1
0
Fork 0
mirror of https://github.com/openclaw/openclaw.git synced 2026-05-05 17:10:20 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix(exec): fail closed when sandbox is unavailable and harden deny followups (#56800)

Browse Source 
* fix(exec): fail closed when sandbox is unavailable and harden deny followups

* docs(changelog): note exec fail-closed fix
This commit is contained in:
scoootscooob
2026-03-28 22:20:49 -07:00
committed by GitHub
parent d5e59621a7
commit 5d81b64343
14 changed files with 116 additions and 52 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
docs/gateway/security/index.md
View File

@@ -191,7 +191,7 @@ If more than one person can DM your bot:
- **Local disk hygiene** (permissions, symlinks, config includes, “synced folder” paths).
- **Plugins** (extensions exist without an explicit allowlist).
- **Policy drift/misconfig** (sandbox docker settings configured but sandbox mode off; ineffective `gateway.nodes.denyCommands` patterns because matching is exact command-name only (for example `system.run`) and does not inspect shell text; dangerous `gateway.nodes.allowCommands` entries; global `tools.profile="minimal"` overridden by per-agent profiles; extension plugin tools reachable under permissive tool policy).
- **Runtime expectation drift** (for example `tools.exec.host="sandbox"` while sandbox mode is off, which runs directly on the gateway host).
- **Runtime expectation drift** (for example `tools.exec.host="sandbox"` while sandbox mode is off, which now fails closed because no sandbox runtime is available).
- **Model hygiene** (warn when configured models look legacy; not a hard block).
If you run `--deep`, OpenClaw also attempts a best-effort live Gateway probe.
@@ -253,8 +253,8 @@ High-signal `checkId` values you will most likely see in real deployments (not e
| `logging.redact_off` | warn | Sensitive values leak to logs/status | `logging.redactSensitive` | yes |
| `sandbox.docker_config_mode_off` | warn | Sandbox Docker config present but inactive | `agents.*.sandbox.mode` | no |
| `sandbox.dangerous_network_mode` | critical | Sandbox Docker network uses `host` or `container:*` namespace-join mode | `agents.*.sandbox.docker.network` | no |
| `tools.exec.host_sandbox_no_sandbox_defaults` | warn | `exec host=sandbox` resolves to host exec when sandbox is off | `tools.exec.host`, `agents.defaults.sandbox.mode` | no |
| `tools.exec.host_sandbox_no_sandbox_agents` | warn | Per-agent `exec host=sandbox` resolves to host exec when sandbox is off | `agents.list[].tools.exec.host`, `agents.list[].sandbox.mode` | no |
| `tools.exec.host_sandbox_no_sandbox_defaults` | warn | `exec host=sandbox` fails closed when sandbox is off | `tools.exec.host`, `agents.defaults.sandbox.mode` | no |
| `tools.exec.host_sandbox_no_sandbox_agents` | warn | Per-agent `exec host=sandbox` fails closed when sandbox is off | `agents.list[].tools.exec.host`, `agents.list[].sandbox.mode` | no |
| `tools.exec.security_full_configured` | warn/critical | Host exec is running with `security="full"` | `tools.exec.security`, `agents.list[].tools.exec.security` | no |
| `tools.exec.auto_allow_skills_enabled` | warn | Exec approvals trust skill bins implicitly | `~/.openclaw/exec-approvals.json` | no |
| `tools.exec.allowlist_interpreter_without_strict_inline_eval` | warn | Interpreter allowlists permit inline eval without forced reapproval | `tools.exec.strictInlineEval`, `agents.list[].tools.exec.strictInlineEval`, exec approvals allowlist | no |
@@ -534,7 +534,7 @@ Even with strong system prompts, **prompt injection is not solved**. System prom
- Prefer mention gating in groups; avoid “always-on” bots in public rooms.
- Treat links, attachments, and pasted instructions as hostile by default.
- Run sensitive tool execution in a sandbox; keep secrets out of the agent’s reachable filesystem.
- Note: sandboxing is opt-in. If sandbox mode is off, exec runs on the gateway host even though tools.exec.host defaults to sandbox, and host exec does not require approvals unless you set host=gateway and configure exec approvals.
- Note: sandboxing is opt-in. If sandbox mode is off, `host=sandbox` fails closed even though tools.exec.host defaults to sandbox. To run on the gateway host, set `host=gateway` and configure exec approvals.
- Limit high-risk tools (`exec`, `browser`, `web_fetch`, `web_search`) to trusted agents or explicit allowlists.
- If you allowlist interpreters (`python`, `node`, `ruby`, `perl`, `php`, `lua`, `osascript`), enable `tools.exec.strictInlineEval` so inline eval forms still need explicit approval.
- **Model choice matters:** older/smaller/legacy models are significantly less robust against prompt injection and tool misuse. For tool-enabled agents, use the strongest latest-generation, instruction-hardened model available.