Loctree is pre-1.0 and moves fast. We support the most recently published
release on each distribution channel (crates.io, npm, Homebrew, the signed
loct.io bundles):
| Channel | Supported |
|---|---|
crates.io (loctree, loctree-mcp) |
latest published version only |
npm (@loctree/loctree) |
latest published version only |
loct.io/install.sh bundle |
latest published version only |
Homebrew (loctree/cli, loctree/mcp) |
latest published version only |
Older release lines do not receive backported security fixes. Upgrading to the latest release is the supported remediation path.
Do not open a public GitHub issue for a suspected vulnerability.
Email support@loctree.com with:
- A description of the issue and its potential impact.
- Steps to reproduce (a minimal repo or command sequence is ideal — loctree runs entirely on local file input, so most issues are reproducible from a small fixture).
- The affected component (
loctree/loctCLI,loctree-mcp,loctree-lsp, the VS Code or JetBrains extension, or the npm/Homebrew packaging) and the version you tested.
If you want encrypted communication, say so in your first message and we will arrange a channel — we do not maintain a published PGP key today.
This is a small team, not a dedicated security function, so we cannot commit to a contractual SLA. In practice:
- We aim to acknowledge reports within 5 business days.
- We will tell you honestly if a fix is going to take longer than that, rather than go silent.
- We will credit reporters in the release notes if you want credit, or keep you anonymous if you prefer.
We ask for coordinated disclosure: please give us a reasonable window to investigate and ship a fix before any public write-up. In turn, we will keep you informed of progress and agree on a disclosure date with you rather than impose one. For a genuinely actively-exploited issue, tell us that explicitly so we can prioritize accordingly.
- Loctree analyzes source code you point it at; it does not execute that source code. Reports about the analyzer mis-parsing malicious/malformed source files as a crash or hang are in scope. Reports assuming loctree executes scanned code are not applicable — it doesn't.
- The MCP/LSP servers accept local stdio/HTTP connections from an editor or agent client. Reports about unauthenticated network exposure, path traversal outside the configured project root, or privilege escalation are in scope.
- Supply-chain concerns (compromised release artifact, dependency
vulnerability with a real exploit path) are in scope — see
make semgrepand.github/workflows/semgrep.ymlfor the automated gate we already run.