moonlight-http 0.1.2

HTTP shadow proxy for comparing candidate service behavior against reference targets.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58

# Moonlight

Moonlight is a Rust and React behavior comparer plus project-check evaluator for validating new implementations against known-good baselines.

## Install

Install the Rust CLI:

```sh
cargo install moonlight-cli --locked
moonlight run --primary 'printf "{\"value\":42}\n"' --candidate 'printf "{\"value\":43}\n"'
```

Run it through npm:

```sh
npx @moritzbrantner/moonlight run \
  --primary 'printf "{\"value\":42}\n"' \
  --candidate 'printf "{\"value\":43}\n"'
```

Run it through Bun:

```sh
bunx @moritzbrantner/moonlight run \
  --primary 'printf "{\"value\":42}\n"' \
  --candidate 'printf "{\"value\":43}\n"'
```

Evaluate a coding-agent patch against an existing project:

```sh
git diff --binary main > agent.patch
moonlight eval run --project moonlight.eval.toml --candidate-patch agent.patch --format markdown
```

## Agent Workflow

For coding-agent goals, create an Agent Goal issue and follow
[`AGENTS.md`](AGENTS.md). Before handoff, run:

```sh
bun run agent:check
```

Moonlight can also evaluate an agent branch or patch against the baseline:

```sh
bun run agent:eval -- --candidate-ref <branch-or-sha>
```

The GitHub Pages site explains the repository layout and shows the latest HTTP and CLI benchmark reports:

<https://moritzbrantner.github.io/moonlight/?page=overview>

For detailed local usage, see [docs/moonlight/README.md](docs/moonlight/README.md).

A local self-dogfood regression harness compares the published CLI against the current source build; see [tests/selfdogfood/README.md](tests/selfdogfood/README.md).