specsync 4.1.0

Bidirectional spec-to-code validation with schema column checking — 11 languages, single binary
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

# .specsync.toml — Alternative to specsync.json
# Place this file in your project root as `.specsync.toml`

# Directory containing spec files (default: "specs")
specs_dir = "specs"

# Source directories to scan (auto-detected if omitted)
source_dirs = ["src", "lib"]

# SQL schema directory for db_tables validation (optional)
# schema_dir = "db/migrations"

# Required markdown sections in each spec
required_sections = [
    "Purpose",
    "Public API",
    "Invariants",
    "Behavioral Examples",
    "Error Cases",
    "Dependencies",
    "Change Log",
]

# Directories to exclude from source scanning
exclude_dirs = ["__tests__", "fixtures"]

# Glob patterns to exclude
exclude_patterns = ["**/__tests__/**", "**/*.test.ts", "**/*.spec.ts"]

# AI provider for `specsync generate --provider <name>`
# Options: auto, claude, anthropic, openai, ollama, copilot, gemini, deepseek, groq, mistral, xai, together
# ai_provider = "auto"

# AI model override (optional)
# ai_model = "gpt-4o"

# Timeout in seconds for AI generation (default: 120)
# ai_timeout = 120