SimCrew SimCrewGive SimCrew a git repo URL. It spins up isolated agent containers — one per persona — each clones your repo and evaluates it like a real developer. Findings come back deduplicated, scored, and ready to act on.
# clone and run git clone git@github.com:mpk-droid/SimCrew.git && cd SimCrew export NVIDIA_API_KEY=nvapi-... # or ANTHROPIC_API_KEY docker compose up
Four steps, no custom harness to write.
Identity, perspective, and constraints — e.g. "backend developer who's never used AI" or "platform engineer evaluating for OpenShift."
Ordered phases like "Read the docs," "Set up locally," "Test the API," "Try deploying."
Provide a git repository URL. That's the only input a run needs.
Each persona clones the repo independently, follows the journey, and reports findings with evidence.
The orchestrator manages the full lifecycle — spin-up, cloning, collection, cleanup.
repo_url ──▶ Orchestrator (FastAPI + Postgres) │ ┌──────────┼──────────┐ │ │ │ Agent: Priya Agent: Sam Agent: Dana ... (container) (container) (container) │ │ │ git clone git clone git clone read docs follow setup read source run commands try to run test edges │ │ │ └────▶ Findings ◀────┘ │ Deduplicate & Score (GREEN / YELLOW / RED)
Every persona runs in its own Docker container, cloning and evaluating the target repo independently — no shared state between runs.
Personas are defined by fields, not raw prompts. SimCrew generates the system prompt, and you review and approve it before it's used.
Journeys are sequences of phases each persona follows independently, from first impressions through deployment.
Findings across personas are deduplicated and scored GREEN / YELLOW / RED, so you see what actually matters first.
Run evaluations from the included UI or drive everything through the REST API — full OpenAPI docs included.
Extend the base agent image to simulate a different OS or toolchain per persona. In active development.
Paired with a 5-phase journey: First Impressions → Setup → Running Locally → Using the Target → Deployment.