enterprise-delivery 0.1.0

package/skills/delivery-lead/SKILL.md

@@ -0,0 +1,221 @@
+---
+name: delivery-lead
+description: "Use when leading AI agents through enterprise delivery work: starting a project, adding or analyzing requirements, deciding tech stack and coding rules, starting implementation tasks, routing agents to the right repo files, or checking completion readiness."
+---
+
+# Delivery Lead
+
+Lead agents through repo-first delivery without adding ceremony. Use this skill when the user asks to start a project, add a requirement, analyze a requirement, choose or review tech stack, update coding rules, start a task, or coordinate completion.
+
+## Core Rule
+
+`docs/enterprise/` is the source of truth. External tools may be linked, but agents must read and update repository artifacts before making delivery-impacting changes.
+
+For uncertain facts, write `unknown` or `inferred`. Do not invent stakeholders, approvals, roadmap commitments, sprint commitments, release readiness, or history.
+
+## Always Read First
+
+For any meaningful work, read these files before proposing or editing:
+
+- `AGENTS.md`: local agent rules and required gates.
+- `docs/enterprise/project-vision.md`: goals, non-goals, principles, forbidden patterns, quality bar, AI collaboration rules.
+- `docs/enterprise/current-state.md`: active stories, changes, failing tests, unresolved bugs, next commands.
+- `docs/enterprise/task-graph.json`: task status, dependencies, story/change links, evidence.
+
+If the work touches architecture, stack, risks, or delivery state, also read:
+
+- `docs/enterprise/architecture-overview.md`
+- `docs/enterprise/test-strategy.md`
+- `docs/enterprise/risk-register.md`
+- `docs/enterprise/decision-log.md`
+- `docs/enterprise/product-backlog.md`
+
+## Workflow Router
+
+### 1. Start Or Onboard A Project
+
+Read:
+
+- `README.md`
+- package/build files such as `package.json`, `pyproject.toml`, `Cargo.toml`, `go.mod`
+- existing docs and tests
+- `AGENTS.md` if present
+
+Update:
+
+- `docs/enterprise/project-charter.md`
+- `docs/enterprise/project-vision.md`
+- `docs/enterprise/architecture-overview.md`
+- `docs/enterprise/test-strategy.md`
+- `docs/enterprise/current-state.md`
+- `docs/enterprise/task-graph.json`
+- `docs/enterprise/risk-register.md`
+- `docs/enterprise/decision-log.md`
+
+Run:
+
+```bash
+node scripts/enterprise-delivery-validate.mjs validate repo
+```
+
+For brownfield projects, create or update `docs/enterprise/onboarding-audit.md` with confirmed, inferred, and unknown facts.
+
+### 2. Add A Requirement
+
+Read:
+
+- `docs/enterprise/project-vision.md`
+- `docs/enterprise/product-backlog.md`
+- relevant existing `docs/enterprise/stories/STORY-*.md`
+- relevant existing `docs/enterprise/changes/CHG-*.md`
+
+Update:
+
+- create or update `docs/enterprise/stories/STORY-<next>-<slug>.md`
+- create or update linked `docs/enterprise/changes/CHG-<next>-<slug>.md`
+- add `linked_changes` in story frontmatter
+- add `related_stories` in change frontmatter
+- update `docs/enterprise/product-backlog.md`
+- update `docs/enterprise/current-state.md`
+
+Required story content:
+
+- Requirement source
+- Business value
+- Acceptance criteria
+- Non-functional criteria
+- Dependencies
+- Docs impact
+- Risk impact
+- Planned test evidence
+
+Keep status `draft` until requirement source and acceptance criteria are explicit.
+
+### 3. Analyze A Requirement
+
+Read:
+
+- target `STORY-*`
+- linked `CHG-*`
+- `docs/enterprise/project-vision.md`
+- `docs/enterprise/architecture-overview.md`
+- `docs/enterprise/test-strategy.md`
+- `docs/enterprise/risk-register.md`
+- `docs/enterprise/decision-log.md`
+
+Do:
+
+- clarify ambiguity as assumptions, questions, or `unknown`
+- split broad work into smaller stories when one story has multiple independent outcomes
+- map each acceptance criterion to implementation tasks and verification commands
+- identify docs, risk, and decision impacts
+
+Update:
+
+- target `STORY-*`
+- linked `CHG-*`
+- `docs/enterprise/task-graph.json`
+- `docs/enterprise/risk-register.md` if risk changes
+- `docs/enterprise/decision-log.md` if architecture, stack, delivery, or governance decisions change
+
+### 4. Analyze Or Choose Tech Stack
+
+Read:
+
+- `docs/enterprise/project-charter.md`
+- `docs/enterprise/project-vision.md`
+- `docs/enterprise/architecture-overview.md`
+- `docs/enterprise/test-strategy.md`
+- package/build files
+- existing source layout
+
+Update:
+
+- `docs/enterprise/architecture-overview.md` with confirmed, inferred, and unknown stack facts
+- `docs/enterprise/decision-log.md` for durable choices
+- `docs/enterprise/test-strategy.md` for chosen verification commands
+- `docs/enterprise/risk-register.md` for stack risks
+
+Decision-log entries should include the decision, reason, and status. Do not mark a tech choice accepted unless the user confirmed it or it is already implemented in the repo.
+
+### 5. Update Coding Rules
+
+Read:
+
+- `AGENTS.md`
+- `docs/enterprise/project-vision.md`
+- `docs/enterprise/architecture-overview.md`
+- `docs/enterprise/test-strategy.md`
+
+Update:
+
+- `AGENTS.md` for operational instructions agents must follow
+- `docs/enterprise/project-vision.md` for durable principles, forbidden patterns, coding philosophy, quality bar, and AI collaboration rules
+- `docs/enterprise/decision-log.md` for rule changes that affect architecture, delivery, release readiness, governance, or team behavior
+
+Keep rules concrete and enforceable. Prefer "Run `npm test` and feature gate before completion" over vague rules like "write high quality code."
+
+### 6. Start A Task
+
+Read:
+
+- `AGENTS.md`
+- `docs/enterprise/project-vision.md`
+- `docs/enterprise/current-state.md`
+- `docs/enterprise/task-graph.json`
+- target `STORY-*`
+- linked `CHG-*`
+
+Before coding:
+
+- confirm the story has acceptance criteria
+- confirm the story and change are bidirectionally linked
+- update or create task entries in `task-graph.json`
+- identify the focused verification commands
+
+Implementation rule:
+
+- use TDD for behavior changes
+- keep edits scoped to the story/change
+- update risk and decision logs when scope changes
+
+### 7. Complete A Task
+
+Read:
+
+- target `STORY-*`
+- linked `CHG-*`
+- `docs/enterprise/task-graph.json`
+- `docs/enterprise/current-state.md`
+
+Update:
+
+- story `Test Evidence` with actual commands and results
+- change `Evidence` with actual commands and results
+- story `Implementation Links`
+- change status to `implemented` or `verified`
+- task status and evidence in `task-graph.json`
+- `current-state.md` for failing tests, unresolved bugs, active work, and next commands
+
+Run:
+
+```bash
+npm test
+node scripts/enterprise-delivery-validate.mjs validate repo
+node scripts/enterprise-delivery-validate.mjs validate feature --story STORY-0001
+```
+
+Only use sprint or release gates when the user is asking about sprint or release readiness.
+
+## Leadership Prompt Pattern
+
+When delegating to an agent, include:
+
+```text
+Read AGENTS.md plus docs/enterprise/project-vision.md, current-state.md, and task-graph.json first.
+Use STORY-____ and CHG-____ as the delivery source of truth.
+Before coding, confirm acceptance criteria and planned evidence.
+During work, update task-graph.json if task status or dependencies change.
+Before completion, run tests, validate repo, validate feature, and record evidence.
+Do not invent unknown facts; mark them unknown or inferred.
+```

package/skills/delivery-planning/SKILL.md

@@ -0,0 +1,27 @@
+---
+name: delivery-planning
+description: Map accepted Enterprise Delivery stories into implementation tasks, planned evidence, documentation updates, risks, and release or sprint traceability before coding starts.
+---
+
+# Delivery Planning
+
+Use this skill after a story and linked change record exist, and before implementation begins.
+
+## Checklist
+
+1. Read the story under `docs/enterprise/stories/` and each linked change under `docs/enterprise/changes/`.
+2. Confirm the story has acceptance criteria, business value, docs impact, risk impact, and planned test evidence.
+3. Break the acceptance criteria into small implementation tasks.
+4. Record each task in `docs/enterprise/task-graph.json` with `story_id`, `change_id`, dependencies, status, and expected evidence.
+5. Update `docs/enterprise/current-state.md` with active story, active change, and next verification commands.
+6. Update `docs/enterprise/risk-register.md` or `docs/enterprise/decision-log.md` when planning reveals a new delivery risk or decision.
+
+## Output
+
+Produce a concise task plan that preserves this chain:
+
+```text
+Requirement -> STORY-* -> CHG-* -> tests/evidence -> sprint/release traceability
+```
+
+Do not invent commitments, stakeholders, approvals, or historical facts. Mark uncertain facts as `unknown` or `inferred`.

package/skills/enterprise-kickoff/SKILL.md

@@ -0,0 +1,25 @@
+---
+name: enterprise-kickoff
+description: Use when adding Enterprise Delivery to a greenfield or brownfield repository, initializing docs/enterprise, auditing current project facts, or establishing baseline agile delivery artifacts
+---
+
+# Enterprise Kickoff
+
+Initialize or audit repo-first enterprise delivery artifacts.
+
+## Rules
+
+- Do not invent project history, stakeholders, roadmap, sprint commitments, release state, or approvals.
+- In brownfield repositories, label facts as confirmed, inferred, or unknown.
+- Create baseline artifacts under `docs/enterprise/`.
+- Create the AI-native state layer under `docs/enterprise/`: `project-vision.md`, `current-state.md`, and `task-graph.json`.
+- Run repository validation before claiming kickoff is complete.
+
+## Workflow
+
+1. Detect whether the repository is greenfield or brownfield by inspecting README, docs, source files, tests, package metadata, and recent commits.
+2. Create missing baseline artifacts from Enterprise Delivery templates, including project vision, current state, and task graph.
+3. For greenfield projects, ask for project charter facts before recording them as confirmed.
+4. For brownfield projects, create `docs/enterprise/onboarding-audit.md` with confirmed, inferred, and unknown facts.
+5. From the target repository, run `node <plugin-install-dir>/scripts/enterprise-delivery-validate.mjs validate repo` or equivalent installed command.
+6. Report validation failures as documentation gaps, not as implementation success.

package/skills/feature-gate/SKILL.md

@@ -0,0 +1,22 @@
+---
+name: feature-gate
+description: Use before claiming a story, feature, or delivery-impacting change is complete
+---
+
+# Feature Gate
+
+Validate story completion evidence before any completion claim.
+
+## Rules
+
+- Run feature validation before moving a story to `done`.
+- Treat validation exit code `1` as a blocker, not a tool failure.
+- Do not claim completion while story, change, acceptance, test, docs, or implementation evidence is missing.
+
+## Workflow
+
+1. Identify the story ID.
+2. Run the installed plugin validator, for example `node <plugin-install-dir>/scripts/enterprise-delivery-validate.mjs validate feature --story <story-id>`, or an equivalent installed command using the identified story ID.
+3. If validation fails, update the story, linked change records, evidence, docs impact, risks, or decisions.
+4. Re-run validation.
+5. Only after validation passes may the story move to `done` or the agent claim the feature complete.

package/skills/release-gate/SKILL.md

@@ -0,0 +1,24 @@
+---
+name: release-gate
+description: Use before claiming a release candidate, release plan, or release package is ready
+---
+
+# Release Gate
+
+Validate release traceability and evidence.
+
+## Rules
+
+- Releases must include release plan, release notes, and traceability matrix.
+- Included stories must pass feature validation.
+- Included changes must be verified before release.
+- Open risks and blockers must be acknowledged.
+
+## Workflow
+
+1. Read `docs/enterprise/releases/<release>/release-plan.md`.
+2. Read `docs/enterprise/releases/<release>/release-notes.md`.
+3. Read `docs/enterprise/releases/<release>/traceability-matrix.md`.
+4. Run the installed plugin validator, for example `node <plugin-install-dir>/scripts/enterprise-delivery-validate.mjs validate release --release <release>`, or an equivalent installed command.
+5. Fix missing traceability, unverified changes, failing stories, risks, blockers, or missing test evidence.
+6. Re-run validation before claiming release readiness.

package/skills/sprint-gate/SKILL.md

@@ -0,0 +1,23 @@
+---
+name: sprint-gate
+description: Use when starting, reviewing, updating, or closing a sprint
+---
+
+# Sprint Gate
+
+Validate sprint delivery state and reporting.
+
+## Rules
+
+- Sprint plans must reference committed story IDs.
+- Done stories inside a sprint must pass feature validation.
+- Blocked stories require a next action; carry-over should be called out in sprint reporting.
+- Risk and decision status must be updated or explicitly marked unchanged.
+
+## Workflow
+
+1. Read `docs/enterprise/sprints/<sprint>/sprint-plan.md`.
+2. Read `docs/enterprise/sprints/<sprint>/status-report.md`.
+3. Run the installed plugin validator, for example `node <plugin-install-dir>/scripts/enterprise-delivery-validate.mjs validate sprint --sprint <sprint>`, or an equivalent installed command.
+4. Fix missing sprint artifacts, story links, change links, blocked next actions, risk updates, or decision updates.
+5. Re-run validation before claiming sprint state is ready, reviewed, or closed.

package/skills/start-task/SKILL.md

@@ -0,0 +1,39 @@
+---
+name: start-task
+description: Use when the user says start task, begin implementation, pick next task, work on STORY-*, implement CHG-*, or asks an agent to begin coding from existing delivery artifacts.
+---
+
+# Start Task
+
+Prepare implementation from the repo source of truth.
+
+## Read
+
+- `AGENTS.md`
+- `docs/enterprise/project-vision.md`
+- `docs/enterprise/current-state.md`
+- `docs/enterprise/task-graph.json`
+- target `docs/enterprise/stories/STORY-*.md`
+- linked `docs/enterprise/changes/CHG-*.md`
+- `docs/enterprise/test-strategy.md`
+- relevant source and test files
+
+## Before Coding
+
+- Confirm the story has acceptance criteria.
+- Confirm story `linked_changes` and change `related_stories` are bidirectional.
+- Identify the task entry in `task-graph.json` or create one.
+- Set task status to `in_progress`.
+- Identify focused test commands.
+- Use TDD for behavior changes.
+
+## Update
+
+- `docs/enterprise/task-graph.json`
+- `docs/enterprise/current-state.md`
+- target `STORY-*` and `CHG-*` if planned evidence, scope, risk, or docs impact changes
+
+## Do Not
+
+- Start implementation when no story/change exists for delivery-impacting work.
+- Broaden scope beyond the story without updating delivery artifacts.

package/skills/story-intake/SKILL.md

@@ -0,0 +1,23 @@
+---
+name: story-intake
+description: Use before starting any feature, bugfix, or delivery-impacting work to create or update a STORY-* artifact and mandatory linked CHG-* record
+---
+
+# Story Intake
+
+Every delivery-impacting task starts with a story and a change record.
+
+## Rules
+
+- Do not begin implementation until the story has a requirement, acceptance criteria, and at least one linked `CHG-*`.
+- Every story must have a linked change record from creation, including draft stories.
+- External tools may be linked but are not the source of truth.
+
+## Workflow
+
+1. Find or create `docs/enterprise/stories/STORY-<next>-<slug>.md`.
+2. Find or create `docs/enterprise/changes/CHG-<next>-<slug>.md`.
+3. Link the change ID in story frontmatter under `linked_changes`.
+4. Link the story ID in change frontmatter under `related_stories`.
+5. Record requirement source, business value, acceptance criteria, docs impact, risk impact, and test evidence plan.
+6. Keep status as `draft` until acceptance criteria and requirement source are explicit.

package/skills/update-coding-rules/SKILL.md

@@ -0,0 +1,32 @@
+---
+name: update-coding-rules
+description: Use when the user asks to add, change, tighten, or review coding rules, AI agent rules, architecture principles, forbidden patterns, quality gates, or repository workflow instructions.
+---
+
+# Update Coding Rules
+
+Update durable rules where agents will actually read them.
+
+## Read
+
+- `AGENTS.md`
+- `docs/enterprise/project-vision.md`
+- `docs/enterprise/architecture-overview.md`
+- `docs/enterprise/test-strategy.md`
+- `docs/enterprise/decision-log.md`
+- `docs/enterprise/current-state.md`
+
+## Update
+
+- `AGENTS.md` for operational instructions agents must follow.
+- `docs/enterprise/project-vision.md` for principles, forbidden patterns, coding philosophy, quality bar, and AI collaboration rules.
+- `docs/enterprise/test-strategy.md` for required verification commands.
+- `docs/enterprise/decision-log.md` for durable decisions.
+- active `STORY-*` and `CHG-*` if this is part of delivery-impacting work.
+
+## Rules
+
+- Make rules concrete and enforceable.
+- Prefer exact commands and file paths.
+- Avoid vague rules like "write clean code" unless paired with observable behavior.
+- Do not conflict with higher-priority user instructions in `AGENTS.md`.

package/src/cli.mjs

@@ -0,0 +1,220 @@
+import path from 'node:path';
+import { initializeEnterprise } from './init.mjs';
+import { installClaudePlaceholder, installCodexPlugin } from './install-codex.mjs';
+import { validateFeature, validateRelease, validateRepo, validateSprint } from './rules.mjs';
+
+export async function runCli(argv, cwd) {
+  const parsed = parseArgs(argv, cwd);
+  if (parsed.error) {
+    console.error(parsed.error);
+    printUsage();
+    return 2;
+  }
+
+  let validationResult;
+  if (parsed.command === 'init') {
+    validationResult = initializeEnterprise(parsed.root);
+  } else if (parsed.command === 'install' && parsed.adapter === 'codex') {
+    validationResult = installCodexPlugin();
+  } else if (parsed.command === 'install' && parsed.adapter === 'claude') {
+    validationResult = installClaudePlaceholder();
+  } else if (parsed.command === 'validate' && parsed.gate === 'repo') {
+    validationResult = validateRepo(parsed.root);
+  } else if (parsed.command === 'validate' && parsed.gate === 'feature') {
+    if (!parsed.story) {
+      console.error('Missing required --story for feature validation.');
+      printUsage();
+      return 2;
+    }
+    validationResult = validateFeature(parsed.root, parsed.story);
+  } else if (parsed.command === 'validate' && parsed.gate === 'sprint') {
+    if (!parsed.sprint) {
+      console.error('Missing required --sprint for sprint validation.');
+      printUsage();
+      return 2;
+    }
+    validationResult = validateSprint(parsed.root, parsed.sprint);
+  } else if (parsed.command === 'validate' && parsed.gate === 'release') {
+    if (!parsed.release) {
+      console.error('Missing required --release for release validation.');
+      printUsage();
+      return 2;
+    }
+    validationResult = validateRelease(parsed.root, parsed.release);
+  } else {
+    console.error(`Unsupported command: ${argv.join(' ')}`);
+    printUsage();
+    return 2;
+  }
+
+  if (parsed.json) {
+    console.log(JSON.stringify(validationResult, null, 2));
+  } else {
+    printHuman(validationResult);
+  }
+
+  return validationResult.status === 'pass' ? 0 : 1;
+}
+
+function parseArgs(argv, cwd) {
+  if (argv.length < 1) {
+    return { error: 'Expected command and gate.' };
+  }
+
+  const [command, gate, ...rest] = argv;
+  const parsed = {
+    command,
+    gate,
+    adapter: gate,
+    root: cwd,
+    story: '',
+    sprint: '',
+    release: '',
+    json: false
+  };
+
+  const optionTokens = command === 'init' ? argv.slice(1) : rest;
+  for (let index = 0; index < optionTokens.length; index += 1) {
+    const token = optionTokens[index];
+    if (token === '--root') {
+      const value = optionTokens[index + 1];
+      if (!value || value.startsWith('--')) {
+        return { error: 'Missing value for --root.' };
+      }
+      parsed.root = path.resolve(cwd, value);
+      index += 1;
+    } else if (token === '--story') {
+      const value = optionTokens[index + 1];
+      if (!value || value.startsWith('--')) {
+        return { error: 'Missing value for --story.' };
+      }
+      parsed.story = value;
+      index += 1;
+    } else if (token === '--sprint') {
+      const value = optionTokens[index + 1];
+      if (!value || value.startsWith('--')) {
+        return { error: 'Missing value for --sprint.' };
+      }
+      parsed.sprint = value;
+      index += 1;
+    } else if (token === '--release') {
+      const value = optionTokens[index + 1];
+      if (!value || value.startsWith('--')) {
+        return { error: 'Missing value for --release.' };
+      }
+      parsed.release = value;
+      index += 1;
+    } else if (token === '--json') {
+      parsed.json = true;
+    } else {
+      return { error: `Unknown argument: ${token}` };
+    }
+  }
+
+  if (!['init', 'validate', 'install'].includes(command)) {
+    return { error: `Unknown command: ${command}` };
+  }
+
+  if (command === 'install') {
+    if (!['codex', 'claude'].includes(parsed.adapter)) {
+      return { error: 'Install adapter must be one of: codex, claude.' };
+    }
+    for (const option of ['story', 'sprint', 'release']) {
+      if (parsed[option]) {
+        return { error: `Option --${option} is not allowed for install.` };
+      }
+    }
+    return parsed;
+  }
+
+  if (command === 'init') {
+    for (const option of ['story', 'sprint', 'release']) {
+      if (parsed[option]) {
+        return { error: `Option --${option} is not allowed for init.` };
+      }
+    }
+    return parsed;
+  }
+
+  const allowedOptionsByGate = {
+    repo: new Set(['root', 'json']),
+    feature: new Set(['root', 'json', 'story']),
+    sprint: new Set(['root', 'json', 'sprint']),
+    release: new Set(['root', 'json', 'release'])
+  };
+  const allowedOptions = allowedOptionsByGate[gate];
+  if (allowedOptions) {
+    for (const option of ['story', 'sprint', 'release']) {
+      if (parsed[option] && !allowedOptions.has(option)) {
+        return { error: `Option --${option} is not allowed for ${gate} validation.` };
+      }
+    }
+  }
+
+  return parsed;
+}
+
+function printHuman(validationResult) {
+  if (validationResult.command === 'install') {
+    if (validationResult.adapter === 'codex') {
+      console.log(`Installed Enterprise Delivery Codex plugin at ${validationResult.pluginDir}`);
+      console.log(`Updated marketplace at ${validationResult.marketplacePath}`);
+      return;
+    }
+
+    console.log(validationResult.message);
+    return;
+  }
+
+  if (validationResult.command === 'init') {
+    console.log(`Initialized Enterprise Delivery at ${validationResult.target}`);
+    if (validationResult.created.length > 0) {
+      console.log('');
+      console.log('Created:');
+      for (const artifact of validationResult.created) {
+        console.log(`- ${artifact}`);
+      }
+    }
+    if (validationResult.skipped.length > 0) {
+      console.log('');
+      for (const artifact of validationResult.skipped) {
+        console.log(`Skipped existing: ${artifact}`);
+      }
+    }
+    return;
+  }
+
+  const label = validationResult.target === validationResult.gate
+    ? validationResult.gate
+    : `${validationResult.gate} ${validationResult.target}`;
+
+  if (validationResult.status === 'pass') {
+    console.log(`PASS ${label}`);
+    return;
+  }
+
+  console.log(`FAIL ${label}`);
+  console.log('');
+  console.log('Missing:');
+  for (const error of validationResult.errors) {
+    console.log(`- ${error.message}`);
+  }
+
+  if (validationResult.nextActions.length > 0) {
+    console.log('');
+    console.log('Next actions:');
+    for (const action of validationResult.nextActions) {
+      console.log(`- ${action}`);
+    }
+  }
+}
+
+function printUsage() {
+  console.error('Usage: enterprise-delivery init [--root PATH] [--json]');
+  console.error('       enterprise-delivery install codex [--json]');
+  console.error('       enterprise-delivery install claude [--json]');
+  console.error('       enterprise-delivery validate repo [--root PATH] [--json]');
+  console.error('       enterprise-delivery validate feature --story STORY-0001 [--root PATH] [--json]');
+  console.error('       enterprise-delivery validate sprint --sprint 2026-W20 [--root PATH] [--json]');
+  console.error('       enterprise-delivery validate release --release 2026.05.0 [--root PATH] [--json]');
+}