llm-browser-testkit 0.11.2

LLM-driven browser test framework — define browser test scenarios in TOML with natural language steps and LLM-powered assertions
Documentation
# Build stage
FROM rust:1.89-alpine AS builder

RUN apk add --no-cache musl-dev gcc make

WORKDIR /app
COPY Cargo.toml Cargo.lock ./
COPY src/ src/
COPY tests/ tests/
COPY examples/ examples/

RUN cargo build --release --all-features

# Runtime stage
FROM alpine:3.23.5

ARG ENABLE_AWS_CLI=false
ARG ENABLE_AZURE_CLI=false

RUN apk add --no-cache chromium chromium-chromedriver ca-certificates

RUN if [ "$ENABLE_AWS_CLI" = "true" ]; then \
        apk add --no-cache aws-cli; \
    fi

RUN if [ "$ENABLE_AZURE_CLI" = "true" ]; then \
        apk add --no-cache python3 py3-pip \
        && pip install --no-cache-dir --break-system-packages azure-cli; \
    fi

ENV CHROME_BIN=/usr/bin/chromium-browser
ENV CHROMEDRIVER=/usr/bin/chromedriver

# Keep the runtime PATH explicit so the llm-browser-testkit binary (and the
# entrypoint that resolves it) stays reachable even if the base image's
# default PATH ever changes.
ENV PATH="/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin"

COPY --from=builder /app/target/release/llm-browser-testkit /usr/local/bin/llm-browser-testkit
COPY default-scenario.toml /default-scenario.toml

EXPOSE 3100
# No ENTRYPOINT: job-container mode (act_runner/GitHub Actions) keeps the
# image entrypoint, and a oneshot binary entrypoint kills the container.
# CMD gives a bare `docker run` an interactive shell; running scenarios is
# still one command: `docker run <image> run /scenario.toml ...`.
CMD ["sh"]