misterdev is pre-1.0. Security fixes land on the latest 0.1.x release.

Please report vulnerabilities privately. Do not open a public issue for a security problem.

Two private channels, either is fine:

• GitHub Security Advisories — open a draft advisory at https://github.com/dcondrey/misterdev/security/advisories/new
• Email — david@writerslogic.com

We aim to acknowledge your report within 48 hours and will keep you updated as we investigate and prepare a fix. Coordinated disclosure is appreciated: please give us a reasonable window to ship a fix before any public disclosure.

Do not include real secrets (API keys, tokens, private keys, customer data) in your report. Redact them, or share a minimal synthetic reproduction instead.

misterdev drives an LLM to edit and verify code, calls external tools, and loads third-party extensions. Reports touching the following are in scope and especially valued:

• Model-generated code execution / sandbox escape — a build, test, or gate command, or model-produced code, escaping its intended execution boundary or running commands it should not.
• Path traversal in the file tools — reads or writes escaping the project root (e.g. ../, absolute paths, symlink tricks) through the editing or file-access tools.
• Secret exposure in logs or gates — API keys, tokens, or environment secrets leaking into logs, gate output, reports, error messages, or LLM prompts.
• MCP server / gateway trust — the agentic MCP tool-use path (including remote gateways) invoking arbitrary or unallowlisted remote tools, or trusting malicious tool responses in a way that compromises the host.
• Plugin / entry-point trust — capabilities discovered via the misterdev.tools, misterdev.gates, and misterdev.targets entry points gaining unexpected privileges or bypassing safety controls.

Reports outside this list are still welcome; these are simply the areas where a finding has the highest impact.