feat(composer): discovery-phase plugin from "I want an app that..." to tracker-loaded stories

Plugin pairs with Symphony — Composer's output (issues in the project's
tracker) is Symphony's input. State-aware /composer command walks the
user through 11 phases: concept, users (with workflow probing for hidden
actors), requirements/NFR/IA, design system, interaction design, flows,
mocks, architecture, stories, roadmap, load. Each phase produces a
reviewable artifact in docs/planning/ (or, for design and interaction,
a project-local skill — see below).

Architecture is intentionally LAST, after the spec is complete. Mocks
are JSX/TSX regardless of final stack and pass a "text feasibility"
audit — every visible string is categorized (static/computed/authored),
computed strings are traced to source data, and any string implying an
unspecified feature halts the commit.

Design system and interaction design are produced as PROJECT-LOCAL
SKILLS in .claude/skills/<slug>-design-system/ and
.claude/skills/<slug>-interaction-design/, not as docs in
docs/planning/. Skills auto-load via description matching whenever UI
work happens; a reference doc gets read once and drifts out of context.
The mocks skill imports tokens directly from the design-system skill
for a single source of truth across mocks and (eventually) real code.

ui-system is currently a PLACEHOLDER — captures the minimum needed for
mocks to proceed with <TBD: ...> markers as a return-pass worklist.
Full interview specification is the next pass.

Council touchpoints baked in: PM/skeptic/end-user after requirements,
end-user/a11y after mocks, full architecture council after architecture,
plus auto-additions (legal-triage on regulated data, privacy-advocate
on PII, finance-controller on money flows).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

composer/templates/users.md

@@ -0,0 +1,55 @@
+# Users & workflows
+
+<!--
+Every actor that touches the system. Humans AND non-humans:
+  - other humans (collaborators, viewers, admins)
+  - AI agents
+  - integrations / webhooks
+  - scheduled jobs
+  - the data / export format itself
+  - future-self
+  - backups
+Workflows extracted by asking "what do you do?" and then "who or what
+is on the other side of each verb?"
+-->
+
+## Primary user
+
+**<Name or role>**
+- **Context**: <what they're doing when they reach for this app>
+- **Job to be done**: <what success looks like for them>
+- **Frequency**: <daily / weekly / occasional>
+
+## Other user-classes
+
+<!--
+List every other actor surfaced by the workflow probe. One block each.
+Non-human classes (agents, integrations, format) get the same treatment
+as humans.
+-->
+
+### <user-class name>
+- **Type**: <human / agent / integration / system / format>
+- **Reads**: <what they read>
+- **Writes**: <what they write, if anything>
+- **Notes**: <anything special about their access pattern or constraints>
+
+## Anti-personas
+
+<!--
+Who this app is NOT for. Naming these saves stories later, because
+"can't we also support X?" gets a clean answer.
+-->
+
+- <type of user/use case explicitly excluded>
+- <another>
+
+## Workflows
+
+<!--
+Verb-keyed list. Each workflow is short. Detailed step-by-step lives
+in flows.md.
+-->
+
+- **<verb>** — <one-line description, who's on each side>
+- **<verb>** — …