claude-opencode-mcp executes AI agents with file and shell access inside a
workspace. We take security reports seriously and appreciate responsible
disclosure.
| Version | Supported |
|---|---|
| 1.x | Yes |
| < 1.0 | No |
Please do not open a public issue for a security problem.
Use GitHub's private vulnerability reporting: https://github.com/putuandy/claude-opencode-mcp/security/advisories/new
Include:
- affected version (
claude-opencode-mcp --version), - OpenCode version (
opencode --version), - a minimal reproduction or proof of concept,
- the impact you believe it has,
- any suggested fix.
We aim to acknowledge reports within 72 hours and to ship a fix or mitigation for confirmed issues as quickly as practical. We will credit reporters in the advisory unless you prefer otherwise.
In scope:
- bypassing workspace validation,
workspace.allowedRoots, or path-hint checks, - escaping the agent permission model (e.g. read-only agents writing files,
git push/git commitbypass,.envprotection bypass), - credential exposure through logs, errors, or network requests,
- command or argument injection in bridge-spawned processes,
- denial of service that hangs or crashes a Claude Code session,
- supply-chain issues in this package's own code or dependencies.
Out of scope (documented behavior, see docs/security.md):
- an agent with shell access can write files through shell commands,
.envprotection applies to file tools, not tobash,- prompt injection from repository content influencing a delegated agent,
- arbitrary code execution caused by a binary you explicitly configure via
opencode.binary/OPENCODE_BIN, - vulnerabilities in OpenCode, Claude Code, or model providers themselves (report those upstream).
The bridge assumes:
- the local user account and the workspace contents are trusted to the degree the user chose to delegate,
- the OpenCode server is bound to
127.0.0.1by default, - provider credentials live in OpenCode or the environment, never in this package's configuration or state.
Defenses: canonical-path workspace validation, allowed-root containment,
path-traversal rejection, an OpenCode permission matrix per agent profile
(no ask rules, auto-rejection of stray permission requests), sensitive-file
patterns, git history protections, bounded inputs/outputs, run timeouts and
cancellation, redacted logs, and 0600 state files.
- Keep
workspace.allowedRootsnarrow when delegating across projects. - Leave
opencode.hostnameat127.0.0.1; binding to a routable address exposes the OpenCode HTTP API to your network (and its token/password, if set, becomes the only barrier). - Review diffs with
get_diffbefore accepting changes. - Prefer read-only agents for untrusted repositories.
- Never commit
.claude-opencode/config.jsonif you put credentials in it (provider credentials belong in OpenCode or environment variables).