llm-sca-tooling 0.2.4

LLM-augmented static code analysis tooling.

Phase 0 Python Package Skeleton

Phase 0 establishes the package surfaces that later implementation phases use. This repository already contains Phases 1-8, so the Phase 0 pass is additive: existing indexing, storage, MCP, SARIF, plugin, and QA modules remain intact.

Current Phase 0 baseline:

• Python floor: 3.12.
• Dependency manager: uv with uv.lock as supply-chain evidence.
• CLI framework: Typer + Rich.
• JSONL writers: orjson.
• Formatting/lint order: isort, black, then ruff.
• Architecture gate: .importlinter plus uv run lint-imports.

Implemented skeleton surfaces:

• llm_sca_tooling.config: typed defaults, JSON/TOML loading, environment overrides, and redacted config output.
• llm_sca_tooling.errors: package-level structured errors.
• llm_sca_tooling.telemetry: logger setup and JSONL trace writer.
• llm_sca_tooling.operations: file-backed run-record writer and budget monitor.
• llm_sca_tooling.governance: permission profiles and policy evaluator.
• llm_sca_tooling.harness: Harness Condition Sheet writer.
• llm_sca_tooling.cli.main: Typer/Rich llm-sca-tooling command entrypoint.
• llm_sca_tooling.plugins.registry.NoOpPlugin: no-op plugin for skeleton registry tests.

The historical evidence-sca CLI remains available for existing graph and MCP commands.

Useful commands:

uv sync --extra dev
llm-sca-tooling --version
llm-sca-tooling version
llm-sca-tooling config show
llm-sca-tooling config validate
llm-sca-tooling harness status
llm-sca-tooling run create demo
make verify
make verify-baseline