Please report security issues privately. Do not open a public issue.

Preferred channel: GitHub's private vulnerability reporting — on this repo's Security tab, click Report a vulnerability. This opens a private draft advisory only the maintainer can see.

Fallback: if private reporting isn't available, email the address listed on the maintainer's GitHub profile with the subject line bitacora: security.

When reporting, include:

• the affected file(s) and a brief reproducer (a minimal handoff body, fixture, or workflow snippet is enough);
• the impact you observed (e.g. validator bypass, agent-instruction injection, workflow privilege escalation);
• any suggested fix, if you have one.

The security surface of this repo is small and specific. The following classes count:

• Shell scripts under plugins/bitacora/scripts/ — validators (validate-ctx.sh), the statusline scripts, and any helper that ships with the plugin. Bugs that let a malformed comment slip past the validator, or that allow command injection from a comment body, qualify.
• Skill SKILL.md files — these are agent-executed instructions. A change that would let a Jira comment's content (read by the agent) hijack agent behavior — classic prompt injection / instruction poisoning — qualifies.
• GitHub Actions workflows under .github/workflows/ — anything enabling workflow-injection, untrusted code execution in the target token's context, or unintended privilege escalation across the issue-gate and label automations.

• The user's own Atlassian / Jira instance, Confluence, or other Atlassian-side configuration.
• Third-party MCP servers the user installs (Atlassian Rovo, Google Drive, etc.) — report those to their respective maintainers.
• Downstream Claude Code installations, the ~/.claude/ configuration, or harness behavior outside this repo's source.
• Anything that requires a maintainer-level GitHub token to exploit.

Only the latest commit on main is supported. There are no LTS branches; security fixes ship forward, not backward.