Skip to content

Security: rvema/shale

Security

SECURITY.md

Security Policy

Reporting a vulnerability

Please use GitHub's private vulnerability reporting: go to the Security tab of this repo and click "Report a vulnerability". Do not open a public issue for anything you believe is exploitable.

You can expect an acknowledgment within a few days. Coordinated disclosure is appreciated; we'll credit reporters in the release notes unless you ask otherwise.

Supported versions

The latest release receives security fixes. There are no maintained older release lines — upgrading is always the fix path (brew upgrade shale).

Security model (what's worth probing)

Shale's security posture is documented in docs/product.md §5–§6. The load-bearing properties, in order of how much we'd like to know if you can break them:

  1. The render workflow never executes PR code. It runs on pull_request_target with a write-capable token and is safe only because it reads everything through the API — there is no checkout. If you find any path by which PR-controlled content reaches execution in that workflow, report it immediately.
  2. All evidence text is treated as attacker-controlled. Everything shale-derived passes a sanitizer before entering the PR card (HTML escaped, @mentions/#refs neutralized, link syntax broken, control characters stripped). A bypass that gets markup, mentions, or links through is a vulnerability.
  3. No network from laptop code paths — capture, finalize, and init never touch the network (enforced by test). A dependency change that breaks this property is a vulnerability.
  4. Secrets must not reach committed evidence. Gitignored paths are dropped, and intent/notes/commands pass secret redaction. A realistic payload that carries a credential into .shale/*.yaml is a vulnerability — please include the (defanged) payload shape.

Explicit non-goals

Evidence is tamper-evident, not tamper-proof — edits after capture are flagged on the card, but wholesale fabrication by a determined author with laptop access is out of scope by design, and the card never claims otherwise. Reports about fabricating evidence wholesale will be closed as working-as-documented.

There aren't any published security advisories