scsh 1.12.0

Scoped Skills Helper — preflight a git repo and run its scoped skills in ephemeral containers.
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
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73

# Built-in workflow: the default "watch a bundle of tasks run" demo, runnable on ANY opened
# directory with zero setup (every param has a default). Three steps on THREE different
# harnesses — the images build first, as their own tracked rows, then the steps run as
# `step k/n` rows in the session browser: claude adds, codex multiplies, and grok folds both
# results into a single plain-English summary.txt artifact copied back beside the results.
description: "Three harnesses, three tracked steps: claude adds A+B, codex multiplies X*Y, grok writes a plain-English summary.txt of both."
params:
  A:
    type: int
    default: "2"
    description: "First addend"
  B:
    type: int
    default: "3"
    description: "Second addend"
  X:
    type: int
    default: "4"
    description: "First factor"
  Y:
    type: int
    default: "5"
    description: "Second factor"
steps:
  add:
    agent:
      harness: claude
      model: sonnet
    inputs:
      A: params.A
      B: params.B
    prompt: |
      Read the integers A and B from the environment and compute their sum. Do nothing
      else — no commits, no network, no other files.
    output:
      sum:
        type: int
  multiply:
    agent:
      harness: codex
      model: gpt-5.5
    inputs:
      X: params.X
      Y: params.Y
    prompt: |
      Read the integers X and Y from the environment and compute their product. Do nothing
      else — no commits, no network, no other files.
    output:
      product:
        type: int
  summarize:
    needs: add, multiply
    agent:
      harness: grok
      model: grok-4.5
    inputs:
      A: params.A
      B: params.B
      X: params.X
      Y: params.Y
      SUM: add.sum
      PRODUCT: multiply.product
    prompt: |
      Two earlier steps computed SUM = A + B and PRODUCT = X * Y; all six values are in the
      environment. Write `summary.txt`: one short plain-English paragraph stating both
      computations and their results in full sentences (for example: "Adding 2 and 3 gives
      5, and multiplying 4 by 5 gives 20."), suitable for a human to read on its own. Put
      the same paragraph in the `summary` result field. Do nothing else — no commits, no
      network, no other files.
    output:
      summary:
        type: string
    artifacts: summary.txt