Analyze an existing (possibly legacy) codebase and write a grounded .specclaw/analysis/codebase-report.md covering tech stack, dependencies, architecture, domain, and risks. Works on any language or stack — Node, .NET, Java, Go, Rust, Python, Delphi/Object Pascal, or none of the above — not just Node/.NET. Read-only…
Produce a C4-model architecture view (L1 System Context → L2 Containers → L3 Components → L4 Code, L4 only where warranted) of an existing/legacy codebase, with a Mermaid flowchart plus grounded prose per level, written to .specclaw/analysis/architecture.md. Works on any language or stack — Node, .NET, Java, Go, Rust…
Design, generate, and record the golden-master harness that proves a rebuild behaves identically to the legacy app it replaces. Default mode ranks seams (pure function / stateful service / HTTP-API / data boundary / UI-excluded), declares each seam's capture layer, audits every seam for non-determinism (clocks…
Synthesize the decisions a brownfield rebuild has already made — decisions.md, module-map.md, architecture.md, rebuild-backlog.md — into .specclaw/analysis/target-architecture.md, the target-side counterpart to architecture.md's legacy view. Produces Mermaid C4 diagrams (one Context, one Container, one Component per…
Create the target application foundation in the NEW (rebuild) repo, before any backlog item is developed — the missing stage between "the rebuild backlog exists" and "/specclaw:propose BL-###". Consumes the architecture the rebuild already decided (decisions.md's SQ/CQ answers plus any accepted ADR in this repo) and…
Turn the extraction signals scattered through .specclaw/analysis/.md (Inference:, Mechanical:, Named Gaps, hedging language, cross-doc conflicts, unexercised code paths) into a numbered, human-answerable question set — .specclaw/analysis/clarifications.md — classified as DECISION, DATA, SCOPE, DEFECT, MECHANICAL…
Produce a domain and functional-spec view of an existing/legacy codebase — business entities, relationships, rules, and enumerations (domain-model.md) plus capabilities, workflows, UI inventory, and named gaps (functional-spec.md) — grounded in parsed forms (.dfm/.xaml), type/const declarations, and validation-routine…
Automatically detects the target application's platform (Web/Desktop/Mobile/Hybrid/Embedded) and stack, dynamically selects the most effective E2E testing framework for it with a stated justification, and generates a Page Object Model plus a runnable E2E test script that asserts structural equality against SpecClaw…
Measure the code quality of an existing/legacy codebase and write a machine-readable .specclaw/analysis/quality.json plus a client-curatable quality-report.md — per-module rollups of cyclomatic complexity, function length, duplication and file length, each classified against configurable thresholds, with every metric…
Read the five .specclaw/analysis/.md documents (codebase-report, architecture, domain-model, functional-spec, module-map) — plus, when present, decisions.md, clarifications.md, and .specclaw/baseline/manifest.json/scenarios.md — and write or refresh an ordered .specclaw/analysis/rebuild-backlog.md: the application…
Golden-master comparison — replays captured legacy fixtures (.specclaw/baseline/manifest.json + fixtures/) against the new app's own behaviour and reports MATCH/DIVERGES/ERROR per fixture, with four mutually exclusive selection scopes (a change's BL item, a single --item BL-### backlog item, a --module MOD-### module…
Show where a brownfield rebuild actually stands — one row per bf- phase (analyze, architecture, domain, clarify, ui, baseline, rebuild-plan, blueprint, bootstrap, replay), what each produced, every open item holding a phase back, and the single next command to run. The bf- counterpart to /specclaw:status, which…
Optional UI-fidelity workstream for a rebuild — extract the legacy app's UI structure and visual theme from its source (.specclaw/ui/ui-inventory.md with a permanent SCR-### per screen, design-tokens.json with permanent TK- token groups, and screenshot-checklist.md as a human capture work order), then record the…
★not rated 12 yesterdayA233 tokens
originalMIT
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: