pipeline-sdk 0.1.0 → 0.1.1

package/README.md

@@ -8,7 +8,7 @@ YAML-driven pipeline orchestration for AI agents.
 
 ## Why?
 
-AI coding agents (Claude Code, Cursor, Codex) execute multi-step workflows -- write code, run tests, deploy -- but lack a structured way to enforce stage gates, track evidence, or coordinate handoffs. pipeline-sdk gives your agent a state machine: define stages in YAML, gate transitions on passing tests, and keep a cryptographic audit trail of what happened.
+AI coding agents (Claude Code, Cursor, Windsurf, and others) execute multi-step workflows -- write code, run tests, deploy -- but lack a structured way to enforce stage gates, track evidence, or coordinate handoffs. pipeline-sdk gives your agent a state machine: define stages in YAML, gate transitions on passing tests, and keep a cryptographic audit trail of what happened.
 
 ## What is this?
 
@@ -82,7 +82,7 @@ The YAML file defines stages and transitions. The state machine enforces the rul
 |----------|-------------|
 | `sdlc-full` | Complete software development lifecycle with review and deploy gates |
 | `static-site` | Web portfolio pipeline with accessibility gates |
-| `zship` | Dual-model development workflow (Claude + Codex) |
+| `dual-model` | Dual-model development pipeline with consensus-based quality gates |
 | `infra-gitops` | Infrastructure deployment with approval gates and rollback |
 
 ## MCP Integration
@@ -131,7 +131,7 @@ console.log(result.valid, result.recordCount);
 
 ```bash
 bun install
-bun test              # 284 tests
+bun test              # Run test suite
 bun run lint          # Biome
 bun run typecheck     # TypeScript
 bun run build         # Compile to binary

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "pipeline-sdk",
-	"version": "0.1.0",
+	"version": "0.1.1",
 	"description": "YAML-driven pipeline orchestration SDK with CLI, daemon, MCP server, and adapter system",
 	"module": "index.ts",
 	"type": "module",
@@ -29,7 +29,6 @@
 	"files": [
 		"src/",
 		"schemas/",
-		"templates/",
 		"README.md",
 		"LICENSE"
 	],

package/src/templates/{zship.yaml → dual-model.yaml}

@@ -1,18 +1,21 @@
-# zship.yaml — Dual-Model Development Pipeline
-# 6-stage pipeline with dual-phase gates (Claude + Codex)
-# Converts the zship process into Pipeline SDK format
+# dual-model.yaml — Dual-Model Development Pipeline
+# Two AI models with consensus-based quality gates across 6 stages.
+#
+# NOTE: Custom gate scripts (e.g. ./tools/gates/*.sh) are user-provided.
+# These paths are examples — replace them with your own review scripts
+# or remove the custom gates to use only built-in gates.
 
 pipeline:
-  id: zship
+  id: dual-model
   version: 1
   description: "Dual-model autonomous development pipeline (BRAINSTORM → PLAN → IMPLEMENT → TEST → REVIEW → SHIP)"
   initial: brainstorm
 
   stages:
     brainstorm:
-      description: "Interrogate requirements via grill-me interrogation"
+      description: "Structured requirements analysis with iterative clarification"
       agent:
-        instructions: "Interrogate requirements relentlessly. Produce approved spec at docs/zship/spec.md."
+        instructions: "Analyze and clarify requirements. Identify ambiguities, missing edge cases, and acceptance criteria. Produce approved spec at docs/spec.md."
         tools: [Read, Grep, Glob, Write, WebSearch]
       on:
         SPEC_APPROVED: { target: plan }
@@ -30,7 +33,7 @@ pipeline:
         SPEC_CHANGES: { target: brainstorm }
 
     implement:
-      description: "Ralph Loop: TDD cycles per task with Gate 1 (test quality) and Gate 2 (impl review)"
+      description: "Iterative TDD cycles (red-green-refactor) per task with Gate 1 (test quality) and Gate 2 (implementation review)"
       agent:
         instructions: "Execute TDD plan. RED → GREEN → Gate 1 review → REFACTOR. After all tasks: Gate 2 full review."
         tools: [Read, Write, Edit, Bash, Grep, Glob]
@@ -53,12 +56,12 @@ pipeline:
         MAX_RETRIES: { target: escalate-implement }
 
     test:
-      description: "E2E testing + smoke testing with Gate 3 (coverage) and Gate 4 (smoke gaps)"
+      description: "E2E testing + smoke testing with Gate 3 (coverage) and Gate 4 (coverage gaps)"
       agent:
         instructions: "Create E2E tests covering spec requirements. Run smoke tests. Gates review coverage and gaps."
         tools: [Read, Write, Edit, Bash, Grep, Glob]
       gates:
-        exit: [e2e-tests-pass, coverage-review, smoke-gap-review]
+        exit: [e2e-tests-pass, coverage-review, coverage-gap-review]
       max_retries: 6
       on:
         STAGE_COMPLETE: { target: review }
@@ -76,12 +79,12 @@ pipeline:
         MAX_RETRIES: { target: escalate-test }
 
     review:
-      description: "Gate 5: Adversarial code review (dual-phase Claude + Codex)"
+      description: "Gate 5: Security and performance review (dual-model consensus)"
       agent:
-        instructions: "Process adversarial findings. Fix security, concurrency, performance issues."
+        instructions: "Review code for security vulnerabilities, race conditions, and performance issues. Fix all findings."
         tools: [Read, Edit, Bash, Grep]
       gates:
-        exit: [adversarial-review]
+        exit: [security-performance-review]
       max_retries: 6
       on:
         STAGE_COMPLETE: { target: ship }
@@ -89,10 +92,10 @@ pipeline:
         TEST_CHANGES: { target: test }
 
     fix-review:
-      description: "Fix adversarial review findings from Gate 5"
+      description: "Fix security and performance review findings from Gate 5"
       max_retries: 3
       agent:
-        instructions: "Fix adversarial findings. Verify fixes via tests. Do not introduce regressions."
+        instructions: "Fix review findings. Verify fixes via tests. Do not introduce regressions."
         tools: [Read, Edit, Bash, Grep]
       on:
         FIX_COMPLETE: { target: review }
@@ -159,7 +162,7 @@ pipeline:
   gates:
     plan-document-exists:
       type: custom
-      script: "test -f docs/zship/spec.md"
+      script: "test -f docs/spec.md"
       expect: { status: pass }
     lint-pass:
       type: builtin
@@ -179,16 +182,16 @@ pipeline:
       script: "./tools/gates/tdd-quality-review.sh"
       protocol: json-stdout
       inputs:
-        mode: dual-phase
-        phase_a: claude
-        phase_b: codex
+        mode: dual-model
+        phase_a: model-a
+        phase_b: model-b
 
     impl-review:
       type: custom
       script: "./tools/gates/impl-review.sh"
       protocol: json-stdout
       inputs:
-        mode: dual-phase
+        mode: dual-model
         focus: architecture
 
     coverage-review:
@@ -196,29 +199,29 @@ pipeline:
       script: "./tools/gates/coverage-review.sh"
       protocol: json-stdout
       inputs:
-        mode: dual-phase
+        mode: dual-model
         focus: test_coverage
 
-    smoke-gap-review:
+    coverage-gap-review:
       type: custom
-      script: "./tools/gates/smoke-gap-review.sh"
+      script: "./tools/gates/coverage-gap-review.sh"
       protocol: json-stdout
       inputs:
-        mode: dual-phase
-        focus: smoke_analysis
+        mode: dual-model
+        focus: coverage_analysis
 
-    adversarial-review:
+    security-performance-review:
       type: custom
-      script: "./tools/gates/adversarial-review.sh"
+      script: "./tools/gates/security-performance-review.sh"
       protocol: json-stdout
       inputs:
-        mode: dual-phase
-        focus: adversarial
+        mode: dual-model
+        focus: security_and_performance
 
     final-quality-gate:
       type: custom
       script: "./tools/gates/final-quality-gate.sh"
       protocol: json-stdout
       inputs:
-        mode: dual-phase
+        mode: dual-model
         focus: final_quality