4bade795fd
Unit (`tests/unit/test_user_schemas.py`): - 4 tests on `UserCreate` (happy path, password min length, email validation, invalid type). - 2 tests on `UserUpdate` (all-optional, password validation when set). - 3 tests on `EngagementMemberCreate` (default `"member"`, explicit role, max-length 40). - 4 tests on `PageQuery` (defaults, offset arithmetic, page_size cap, page lower bound). Integration (`tests/integration/test_user_mgmt_e2e.py`, marked `integration`): - The critical MA6-in-practice flow: rt_lead creates rt_operator, assigns to engagement A, the operator signs in, lists engagements and sees only A, `GET /engagements/B` returns 404 (anti-leak), `GET /engagements/B/members` returns 404 too, `/engagements/A/members` is reachable, `GET /users` is forbidden for the operator. - `USER_MANAGE` gate: anonymous → 401, operator session → 403, lead session → 200. - 409 `email_taken` on duplicate `POST /users`. - `/audit/log` is lead-only, paginates with `page_size`, filters by `?action=`. - Disabling a user blocks subsequent logins (same uniform `invalid_credentials` envelope as for bad passwords — no enumeration leak of "this account was disabled"). 74 unit tests pass (61 sprint 1 + 13 sprint 2); integration tests run on the testcontainers Postgres fixture in CI.
Mimic
Internal BAS (Breach & Attack Simulation) platform for the Red Team. Replays TTPs from engagement journals or an internal ATT&CK library against client infrastructure through VPN/relay, in white-glove coordination with the SOC.
Output: a coverage report mapped to MITRE ATT&CK — measurable, reproducible, archived.
Status
ready-with-prereqs — spec frozen on 2026-05-19, 23 review patches integrated.
Code start blocked on:
- PR1 — Mythic API documentation + pinned version (lead RT)
- PR2 — Internal C2 interface spec + journal export example (internal C2 team)
- PR3 — RT graphic charter for the PDF report (lead RT)
While PR1/PR2/PR3 are open, sprint 0 focuses on the unblocked skeleton.
Spec
The authoritative spec lives in the RT-SecondBrain vault:
Projects/Mimic — Spec.md. Do not duplicate it here.
In-repo documentation:
CHANGELOG.md— chronological log of features, decisions, rollbacks.tasks/spec-decisions.md— implementation arbitrations on top of the spec.tasks/todo.md— current sprint backlog.
Stack (frozen)
- Backend Python 3.12 / Flask / Flask-SocketIO / SQLAlchemy 2 / Pydantic 2 / Alembic / WeasyPrint / pytest + testcontainers / ruff / mypy strict
- Frontend TypeScript / React 18+ / Vite / Tailwind 4 / TanStack Query 5 / Recharts / Playwright
- Storage Postgres (prod) / SQLite (pure-logic unit tests) / testcontainers Postgres (audit log, RBAC, write-only role — incl. unit tests of Postgres-specific behavior, per H38)
- Deploy Docker images + Ansible deployment playbook (per D-010). Reverse proxy (Caddy + TLS + IP allowlist) handled by existing RT infrastructure, out of Mimic scope (D-007).
Layout
mimic/
├── backend/ # Flask app, connectors, orchestrator, reporting, CLI
├── frontend/ # Vite + React app
├── docs/ # Architecture notes, ADRs, deployment
└── tasks/ # Sprint backlog, decisions, lessons
Conventions
- Branches:
feature/<scope>,fix/<scope>,docs/<scope>,chore/<scope>. Long-lived:main. - Commits: Conventional Commits (
feat:,fix:,chore:,docs:,test:,refactor:). - PRs: each branch → review (
code-reviewer) → team-lead merges. No direct push tomain.
Build & run
make targets land at the end of sprint 0. For now the repo is skeleton-only.
Licensing
Internal — proprietary, RT use only. Do not redistribute.