Add SDLC workflow + team registry on top of #20 (closes #19)

.github/workflows/test-sdlc-workflow.yml

@@ -0,0 +1,68 @@
+name: Test SDLC Workflow
+
+on:
+  pull_request:
+    paths:
+      - 'agent-context/sdlc-workflow.sh'
+      - 'agent-context/sdlc-pointer.sh'
+      - 'agent-context/README.md'
+      - 'scripts/start-agent-session.sh'
+      - 'scripts/capture-session-memory.sh'
+      - 'scripts/init-project.sh'
+      - 'scripts/upgrade-project.sh'
+      - 'scripts/verify-project-install.sh'
+      - 'scripts/sdlc.sh'
+      - 'scripts/validate-command-adapters.sh'
+      - 'templates/cursor/sdlc-spdd-whereami.md'
+      - 'templates/copilot/prompts/sdlc-spdd-whereami.prompt.md'
+      - 'templates/claude/commands/sdlc-spdd-whereami.md'
+      - 'templates/cursor/rules/sdlc-spdd.mdc'
+      - 'templates/copilot/copilot-instructions.md'
+      - 'templates/claude/CLAUDE.md'
+      - 'tests/test-sdlc-workflow.sh'
+      - 'tests/test-sdlc-pointer.sh'
+      - '.github/workflows/test-sdlc-workflow.yml'
+  push:
+    branches: [main]
+    paths:
+      - 'agent-context/sdlc-workflow.sh'
+      - 'agent-context/sdlc-pointer.sh'
+      - 'agent-context/README.md'
+      - 'scripts/start-agent-session.sh'
+      - 'scripts/capture-session-memory.sh'
+      - 'scripts/init-project.sh'
+      - 'scripts/upgrade-project.sh'
+      - 'scripts/verify-project-install.sh'
+      - 'scripts/sdlc.sh'
+      - 'scripts/validate-command-adapters.sh'
+      - 'templates/cursor/sdlc-spdd-whereami.md'
+      - 'templates/copilot/prompts/sdlc-spdd-whereami.prompt.md'
+      - 'templates/claude/commands/sdlc-spdd-whereami.md'
+      - 'templates/cursor/rules/sdlc-spdd.mdc'
+      - 'templates/copilot/copilot-instructions.md'
+      - 'templates/claude/CLAUDE.md'
+      - 'tests/test-sdlc-workflow.sh'
+      - 'tests/test-sdlc-pointer.sh'
+      - '.github/workflows/test-sdlc-workflow.yml'
+
+jobs:
+  test-sdlc-workflow:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Check shell syntax
+        run: |
+          bash -n agent-context/sdlc-pointer.sh
+          bash -n agent-context/sdlc-workflow.sh
+          bash -n scripts/sdlc.sh
+          bash -n tests/test-sdlc-workflow.sh
+          bash -n scripts/start-agent-session.sh
+          bash -n scripts/capture-session-memory.sh
+          bash -n scripts/validate-command-adapters.sh
+
+      - name: Run SDLC pointer regression tests
+        run: ./tests/test-sdlc-pointer.sh
+
+      - name: Run SDLC workflow regression tests
+        run: ./tests/test-sdlc-workflow.sh

agent-context/README.md

@@ -54,6 +54,9 @@ Each work item also has a canonical canvas under `spdd/canvas/`. Keep both copie
 
 ## SDLC Pointer (current chore/task)
 
+**Quick start:** `./scripts/sdlc.sh` (or `./scripts/sdlc.sh next`) shows what to do now.
+In chat: `/sdlc-spdd-whereami`.
+
 Agents can drift onto the wrong Work ID when several chores are open. The pointer
 manager keeps a single active chore in `.sdlc/pointer` (local state; not committed)
 and provides guarded wrappers so commands refuse to run against a stale pointer.
@@ -77,6 +80,100 @@ sdlc_init
 
 `start-agent-session.sh` sets the pointer automatically when `--work-id` is provided.
 
+## SDLC Workflow (phase + gate tracking)
+
+**Short commands** (installed at `scripts/sdlc-spdd/sdlc.sh`; orchestrator repo: `scripts/sdlc.sh`):
+
+```bash
+./scripts/sdlc.sh              # what to do now (default)
+./scripts/sdlc.sh status       # full dashboard (auto-syncs)
+./scripts/sdlc.sh start        # open session brief at current phase
+./scripts/sdlc.sh resume FEAT-001-order-status-api
+./scripts/sdlc.sh advance
+./scripts/sdlc.sh shelf --reason "blocked"
+./scripts/sdlc.sh capture --summary "finished T02"   # pointer-guarded
+```
+
+In **code** phase, the next canvas operation (`T01`, `T02`, …) is inferred automatically from the REASONS Canvas.
+
+The workflow manager builds on the pointer to answer **where am I?**, **what is next?**,
+and **how do I shelf or resume work?** State lives under `.sdlc/workflows/` (local,
+gitignored). Committed artifacts (`progress-log.md`, canvas, reviews) remain the audit trail;
+run `sync` to reconcile workflow state from those files.
+
+```bash
+# Where am I on the current task?
+./agent-context/sdlc-workflow.sh status
+
+# Pick up a shelved task (auto-shelves the current pointer if different)
+./agent-context/sdlc-workflow.sh resume FEAT-001-order-status-api
+
+# Resume at a specific phase (e.g. after intentionally skipping ahead)
+./agent-context/sdlc-workflow.sh resume FEAT-001-order-status-api --phase code
+
+# Move to the next phase after finishing a step
+./agent-context/sdlc-workflow.sh advance
+
+# Jump ahead to a later phase
+./agent-context/sdlc-workflow.sh advance --to review
+
+# Skip a phase with a recorded reason
+./agent-context/sdlc-workflow.sh skip api-test --reason "no HTTP surface"
+
+# Park current work and clear the pointer
+./agent-context/sdlc-workflow.sh shelf --reason "blocked on dependency"
+
+# Re-read canvas, progress log, and session brief into workflow state
+./agent-context/sdlc-workflow.sh sync
+
+# List shelved work ids
+./agent-context/sdlc-workflow.sh list-shelved
+```
+
+`start-agent-session.sh` and `capture-session-memory.sh` update workflow timestamps
+automatically. After shelving, run `resume <WORK-ID>` then `start-agent-session.sh`
+with the suggested phase to sync back into the chat workflow.
+
+## Team Work ID sharing
+
+Local pointer (`.sdlc/`) is private to your machine. **Team coordination** uses the
+committed file `agent-context/work-registry.tsv` — commit it after claim/release so
+teammates see who is on which Work ID, phase, and operation.
+
+```bash
+./scripts/sdlc.sh list-work              # discover Work IDs in the repo
+./scripts/sdlc.sh team                   # team registry + your pointer
+./scripts/sdlc.sh claim FEAT-001-alpha   # resume + register as active owner
+./scripts/sdlc.sh release --reason "handoff to QA"
+./scripts/sdlc.sh resume OTHER-ID --force   # take over if teammate left stale claim
+```
+
+Set `SDLC_USER="Jane"` to label registry rows. Set `SDLC_NO_TEAM_REGISTRY=1` to opt out.
+
+**Stale claims:** active rows older than `SDLC_TEAM_STALE_DAYS` (default 7) show `[STALE>Nd]` in
+`team` / `list-work`. Stale claims warn but do not block; non-stale claims block until `--force`.
+
+**Done status:** canvases with `## Final Status` → `Status: Complete` are marked `done` when you run
+`team`, `list-work`, or `sync-team`.
+
+**Branch / PR / Jira linking** (stored in the `note` column):
+
+```bash
+./scripts/sdlc.sh claim FEAT-001 --branch cursor/feat-001 --pr "#21" --jira "PROJ-123"
+# auto-detects current git branch on claim (disable: SDLC_TEAM_AUTO_BRANCH=0)
+# auto-reads Jira Key from requirements/milestones/<WORK-ID>.md ## Jira (disable: SDLC_TEAM_AUTO_JIRA=0)
+```
+
+Jira **draft syntax** for issue creation lives in `requirements/milestones/<WORK-ID>.md` under
+`## Jira`. Set `- Key: ABC-123` after create; `list-work` shows `jira:ABC-123` or `jira draft`.
+
+**Notifications:** copy `agent-context/hooks/notify-team-registry.example.sh` and set:
+
+```bash
+export SDLC_TEAM_REGISTRY_HOOK=./agent-context/hooks/notify-team-registry.sh
+export SDLC_TEAM_SLACK_WEBHOOK=https://hooks.slack.com/services/...
+```
+
 ## Session Persistence
 
 Use scripts to keep agent sessions durable across chat boundaries: