If you believe you have found a security issue in PDD, please report it privately. Do not open a public GitHub issue for sensitive reports.

Preferred channel:

• Open a private security advisory on the repository: github.com/blpsoares/parity-driven-development/security/advisories/new

Alternatively, contact the maintainers directly through the repository's GitHub profile (@blpsoares) rather than filing a public issue.

Please include enough detail to reproduce the problem: affected file or command, steps to trigger it, and the impact you observed. We will acknowledge the report and work with you on a fix and coordinated disclosure.

PDD is intentionally small in surface area, so please calibrate expectations:

• It is a documentation/command framework. The core is a set of plain markdown skills (skills/*/SKILL.md) and the .audit/ method — files such as BOOTSTRAP.md, findings/, and coverage.md that live in your project.
• The optional pdd CLI is a local, read-only dashboard. It reads the .audit/ files in your working tree and renders their state. It runs no server, exposes no network service, and handles no credentials or secrets.
• No network services, no secrets handling. PDD does not collect telemetry, make outbound calls of its own, or store authentication material.

Because of this, the realistic security surface is limited to local behavior: the CLI parsing untrusted .audit/ files, and the commands your agent runs on your behalf inside your own repository.

PDD's central safety guarantee is procedural, and it is enforced by the command definitions themselves:

• The AI never authors commits.
• push and gh pr create happen only after an explicit human "yes" in the same session.
• Merge is 100% human, and only after QA approves.

If you find a way to make a PDD command bypass one of these gates — for example, a skill that commits, pushes, or merges without the required human approval — we consider that a security-relevant defect and would like to hear about it through the private channel above.

PDD is distributed as a plugin/framework installed per project. Security fixes are applied to the latest version on the main branch; please update to the newest release before reporting an issue you cannot reproduce there.