empirical-sdd 0.20.3 → 0.22.0

package/README.md

@@ -1,74 +1,119 @@
 # Empirical SDD
 
-Agent-neutral, resumable spec-driven development for 73 global agent-skill
-targets, with verified runtime guidance for Codex, Claude Code, Cursor, Gemini
-CLI, Windsurf, and MCP clients.
+Agent-neutral, resumable spec-driven development with deterministic routing,
+immutable evidence, safe cross-worktree integration, and six agent-native
+skills. Empirical installs across 73 global agent targets and provides verified
+guidance for Codex, Claude Code, Cursor, Gemini CLI, Windsurf, and MCP clients.
 
-> Empirical is alpha software. It supports one active feature per checkout and
-> uses Git worktrees for isolated parallel work.
+> Empirical 0.22 is alpha software and intentionally introduces the breaking
+> Schema 5 protocol. One checkout selects at most one active feature; linked Git
+> worktrees isolate parallel work.
 
 ## Install
 
 | Command | Purpose |
 | --- | --- |
-| `npm install -g empirical-sdd` | Install the Empirical CLI globally. |
-| `empirical install` | Choose coding agents and install their five Empirical skills. |
-| `empirical update` | Upgrade Empirical and refresh the installed skills. |
+| `npm install -g empirical-sdd` | Install the CLI. Node.js 22 or newer is required. |
+| `empirical install` | Choose coding agents and install all six skills. |
+| `empirical update` | Upgrade Empirical and reconcile installed skills. |
+| `empirical uninstall` | Remove managed global skills, owned selection metadata, then the global package. |
 
-Restart or reload each selected agent after installation.
+The installer uses a pinned local catalog. It remembers exact target IDs,
+deduplicates shared skill roots, and performs no runtime network fetch or `npx`
+execution. Automation can use repeatable `--agent`/`-a`, `--all`, `--yes`, and
+`--json` options.
 
-`empirical install` uses a searchable, scrollable local catalog pinned to a
-reviewed `vercel-labs/skills` release. It remembers the exact selected ids,
-deduplicates shared skill roots, and never runs `npx` or uses the network. Type
-to search by agent name, id, or destination; use arrows, Space, Enter, Escape,
-or Ctrl-C. Automation can use repeatable `--agent`/`-a`, `--all`, `--yes`, and
-`--json`; the legacy ids `claude` and `gemini` remain aliases.
+Uninstall is fail-closed: interactive use shows the exact scope and defaults to
+cancel, while automation must use `empirical uninstall --yes` (optionally with
+`--json`). It removes only marker-owned global artifacts. Project `.empirical`
+history, evidence, and repository MCP/agent configuration are always preserved;
+unmanaged or unsafe global paths are also preserved and reported.
 
-## Use
+### Difference from the previous command surface
 
-These are coding-agent skills, not shell commands:
+| Previous 0.21 behavior | Updated behavior |
+| --- | --- |
+| Public lifecycle exposed Install and Update only. | Public lifecycle also exposes `empirical uninstall`. |
+| Removing six agent skills required manual per-agent cleanup. | One command scans every unique catalog root and removes only marker-owned skills. |
+| Selection metadata had to be found manually. | Valid Empirical-owned metadata is removed; invalid or user-owned metadata is preserved. |
+| Package removal was a separate manual npm command. | Confirmed uninstall runs exact `npm uninstall -g empirical-sdd` last. |
+| Project preservation was implicit. | Help, confirmation, human output, and JSON explicitly report preserved project state. |
+
+## Skills
 
-| Agent | Automatic | Init | Spec | Socratic | Loop |
-| --- | --- | --- | --- | --- | --- |
-| Codex | `$empirical` | `$empirical-init` | `$empirical-spec` | `$empirical-socratic` | `$empirical-loop` |
-| Claude Code | `/empirical` | `/empirical-init` | `/empirical-spec` | `/empirical-socratic` | `/empirical-loop` |
-| Cursor | `empirical` | `empirical-init` | `empirical-spec` | `empirical-socratic` | `empirical-loop` |
-| Gemini CLI | `empirical` | `empirical-init` | `empirical-spec` | `empirical-socratic` | `empirical-loop` |
-| Windsurf | `@empirical` | `@empirical-init` | `@empirical-spec` | `@empirical-socratic` | `@empirical-loop` |
+These are coding-agent skills, not public shell workflow commands.
 
 | Skill | Purpose |
 | --- | --- |
-| `empirical <request>` | Initialize if needed, select the workflow, and run the request through review. |
-| `empirical-init` | Initialize or repair repository setup without starting work. |
-| `empirical-spec <request>` | Draft a concrete specification and stop for approval. |
-| `empirical-socratic <idea>` | Refine an idea through a five-pass interview, then draft a specification. |
-| `empirical-loop` | Resume the active approved specification and drive it to completion. |
-
-On first `empirical-init`, the agent shows the complete recommended setup before
-writing anything. Apply the defaults, customize Verification, Parallel work,
-and Decisions one section at a time, or cancel. Repair shows current values and
-keeps them unless edited. Verification independently stores acceptance-test,
-UI-browser, UI-screenshot, and code-review gates; all default on. Turning off
-criterion evidence makes its UI sub-policies inactive without erasing them, and
-does not turn off code review.
-
-Broad catalog membership means an agent can read global `SKILL.md` files. MCP
-bridging and executable handoff remain separately verified capabilities; the
-installer does not infer either one.
+| `empirical <request>` | Initialize, resume, route, and run in normal mode. |
+| `empirical-init` | Initialize or repair repository context without starting work. |
+| `empirical-spec <request>` | Draft a concrete Complex contract and stop for approval. |
+| `empirical-socratic <idea>` | Run the durable five-pass interview, then draft the contract. |
+| `empirical-loop` | Resume the selected feature from its exact current revision. |
+| `empirical-yolo <request>` | Run autonomously to an explicit safe completion ceiling. |
+
+Native invocation examples include `$empirical` in Codex, `/empirical` in
+Claude Code, and `@empirical` in Windsurf. Reload an agent after installation.
+
+## Trust model
+
+Requests are classified into deterministic risk floors: contract-neutral,
+behavioral, sensitive, migration, integration, delivery, or publication. Fast
+is available only to contract-neutral work. Every higher floor uses Complex;
+wording a risky request as “quick” cannot demote it.
+
+Normal mode asks only when a material product choice or permission is missing.
+YOLO persists bounded standing authorization and asks only for genuine blockers
+before its authorized ceiling. It never bypasses host permissions or branch
+protection, force-writes Git, extracts credentials, deletes real worktrees or
+branches, replaces immutable releases, or infers publication.
+
+Fast is contract-neutral and ends at verified. Complex records an impact
+manifest and proceeds through Specify, Design, Plan, Implement, Verify, Review,
+and independent Integrate. An authorized delivery may continue through two
+protected GitHub PRs. Publication is always a separate explicit operation bound
+to an exact version, commit, tag, and dist-tag.
+
+Evidence is not a caller-supplied boolean. Empirical either executes a Policy v2
+command or collects a content-addressed artifact, then writes an immutable
+receipt tied to criteria, source state, and provenance. Completion reports only
+the highest proven level: implemented, verified, integrated, delivered, or
+published.
+
+## Repository model
+
+Schema 5 stores strict Policy v2 configuration, Manifest v2 knowledge
+fingerprints, impact manifests, receipts, Git-common-dir capability claims, and
+hash-chained per-feature journals. Terminal journals compact transactionally to
+a verified snapshot boundary. `empirical_doctor` diagnoses schema, journal,
+lock, claim, toolchain, policy, knowledge, evidence, worktree, and delivery
+state without mutating the repository.
+
+Schema 4 repositories migrate atomically on the first mutating 0.22 operation.
+The migration validates a complete candidate tree before promotion and retains
+a recovery receipt. Earlier schemas must first be upgraded to Schema 4 with the
+version that created them.
 
 ## Development
 
-Requires Node.js 20+ and Bun.
+Development requires Node.js 22+ and Bun. CI covers Node 22, 24, and 26.
 
 | Command | Purpose |
 | --- | --- |
-| `bun install` | Install development dependencies. |
-| `bun run ci` | Run type checking, tests, build smoke tests, and package tests. |
+| `bun install` | Install dependencies. |
+| `bun run check` | Type-check the source. |
+| `bun run test` | Run the test suite. |
+| `bun run test:coverage` | Enforce aggregate and per-module coverage gates. |
+| `bun run test:dist` | Build and smoke-test the packaged CLI and MCP server. |
+| `bun run test:package` | Install and import the generated npm package. |
+| `bun run ci` | Run every required gate. |
+
+The package exposes only `.`, `./protocol`, `./mcp`, and `./integrations`.
 
 ## Documentation
 
-[Architecture](docs/architecture.md) · [Demos](docs/demo.md) ·
-[MCP](docs/mcp.md) · [Security](docs/security.md) ·
+[Protocol](docs/protocol.md) · [Architecture](docs/architecture.md) ·
+[MCP](docs/mcp.md) · [Demo](docs/demo.md) · [Security](docs/security.md) ·
 [Migration](docs/migration-v1.md)
 
 ## License

package/dist/branding.d.ts

@@ -0,0 +1,15 @@
+export declare const BRAND_RED: "#F43737";
+export declare const BRAND_YELLOW: "#FFCD15";
+export declare const BRAND_BLUE: "#4A5CFF";
+export interface BrandBannerOptions {
+    version: string;
+    columns?: number;
+    color?: boolean;
+}
+export interface BrandBannerOutput {
+    isTTY?: boolean;
+    columns?: number;
+}
+export type BrandBannerEnvironment = Readonly<Record<string, string | undefined>>;
+export declare function renderBrandBanner(options: BrandBannerOptions): string;
+export declare function renderBrandBannerForOutput(version: string, output?: BrandBannerOutput, environment?: BrandBannerEnvironment): string;