vibe-action 0.2.2

Command router — execute shell commands and LLM prompts via simple YAML actions.
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

version: 0.0.1

name: review
about: Critically analyze code for bugs and flaws using a powerful model
notify: true

# IDE plugin integration: maps argument names to IDE contexts
api:
  # Target for the final result (replace | clipboard | dialog)
  output: dialog
  # Source for the {query} tag
  input: query

actions:
  # Read input from query (prioritize file path if valid, otherwise use raw text)
  - tag: tag_content
    run: cmd
    expect: string
    check: .+
    action:
      - when: '{query|file_path|empty:not}'
        then: cat "{query|file_path|load}"
      - when: '{query|empty:not}'
        then: echo "{query}"
      - when: 'true'
        then: echo "ERROR - No input provided"

  # Analyze code for bugs and generate technical recommendations in the target language
  - tag: tag_review
    run: large
    expect: string
    action: |
      [Task]
      Analyze the provided code for logic flaws, concurrency bugs, memory leaks, and performance bottlenecks.

      [Rules]
      - Write strictly in the language specified in [User]. Keep feedback concise, blunt, and highly technical.
      - If the code has no flaws, output exactly one word: PERFECT.
      - Format each issue as a clear block: standard bullet point for the description, followed by a new line with a prefix "  Fix: " and the exact 1-line code patch.
      - Do NOT write explanation inside the fix line, provide ONLY the clean code patch.
      - Language: {system_language}

      [Code to Review]
      {tag_content}

  # Copy result to clipboard and output
  - tag: tag_clipboard
    run: value
    expect: string
    action: '{tag_review|clipboard}'