Skip to content
Fencer Intel
medium 6.3

GHSA-j425-whc4-4jgc json

OpenClaw's `system.run` env override filtering allowed dangerous helper-command pivots

published 8mo ago modified 8mo ago

cvss
6.3
version 3.1, GitHub
cisa kev
Not listed
not in CISA's catalog
epss
—
no score, EPSS covers CVEs only
public exploits
0
none known

Summary

### Summary `system.run` env override sanitization allowed dangerous override-only helper-command pivots to reach subprocesses. A caller who could invoke `system.run` with `env` overrides could bypass allowlist/approval intent by steering an allowlisted tool through helper-command or config-loading environment variables such as `GIT_SSH_COMMAND`, editor/pager hooks, and `GIT_CONFIG_*` / `NPM_CONFIG_*`.

### Affected Packages / Versions - Package: `openclaw` (npm) - Latest published vulnerable version: `2026.3.2` - Affected range: `<= 2026.3.2` - Patched in: `2026.3.7`

### Details Before the fix, `src/infra/host-env-security.ts` blocked only a narrow set of override-only environment variables. Dangerous request-scoped overrides such as `GIT_SSH_COMMAND` and prefix families such as `GIT_CONFIG_*` and `NPM_CONFIG_*` could still survive `sanitizeSystemRunEnvOverrides(...)` / `sanitizeHostExecEnv(...)` and reach the spawned process.

That mattered for `system.run` allowlist and approval flows because approval evaluation was tied to the reviewed binary/argv, while the launched process could still inherit attacker-controlled env overrides that changed helper-command execution or config resolution. For allowlisted tools such as `git`, this allowed behavior outside the reviewed command semantics.

The fix extends the shared TypeScript and macOS policy to block dangerous override-only exact keys and prefixes while preserving trusted inherited base-environment behavior.

### Impact This is a real protection-bypass issue, but exploitation requires an already tool-enabled caller who can invoke `system.run` and supply `env` overrides. In affected deployments, that caller could bypass allowlist/approval intent and trigger helper-command execution or config-loading behavior that is not represented by the approved command line. Maintainer severity is set to medium because the bug still requires that existing execution capability; the vulnerability is the mismatch between reviewed command semantics and the actual spawned-process behavior.

### Fix Commit(s) - `e27bbe4982439da6864160fd1b66445058f74801`

### Release Process Note npm `2026.3.7` was published on March 8, 2026. This advisory is fixed in the released package.

Thanks @tdjackey and @SnailSploit for reporting.

Patch / fix

Is GHSA-j425-whc4-4jgc in your dependencies? Fencer shows which of your repositories are affected as soon as an advisory lands, and opens the fix as a pull request.

Check with Fencer

CVSS scores

assigned by version score vector
GitHub 3.1 6.3 CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:L

Affected packages

ecosystem package affected fixed in
npm openclaw
<2026.3.7
2026.3.7

Classification

Weakness
  • CWE-15 External Control of System or Configuration Setting
  • CWE-639 Authorization Bypass Through User-Controlled Key

References

Sources

Fencer Intel keeps one entry per vulnerability: the CVE id wins, then GHSA, and records that alias each other are merged, except that two CVEs always keep separate entries. The same data is available as JSON.

The maintainers of openclaw/openclaw published their own advisory on GitHub. Reviewed sources take precedence over it.