feat(skills): ship default end-user skill pack (bundled v5)

Install the model-invoked workflow catalog on fresh install and upgrade:
debug/test/review/plan/implement and related end-user skills, power
explicit-only batch/dependency-update/release, and docx/pptx/xlsx with
documents/presentations/spreadsheets compatibility aliases.

Non-destructive upgrade rules: never overwrite a skill body that no
longer matches the shipped asset; safely retire unchanged
v4-best-practices via exact body match; keep Feishu optional rather
than universal.

Closes #4691

Signed-off-by: Hunter B <hmbown@gmail.com>

Author: Hmbown

crates/tui/assets/skills/batch/SKILL.md

@@ -0,0 +1,22 @@
+---
+name: batch
+description: Break a large, parallelizable goal into bounded work units, coordinate existing agent/worktree machinery, integrate, and verify. Explicit-only.
+invocation: explicit-only
+---
+
+# Batch
+
+## Invocation
+Explicit-only. Do not start from ambient wording alone.
+
+## When to use
+Use for large parallelizable goals that need coordinated sub-work.
+
+## Non-goals
+- Do not invent a second orchestration engine; reuse delegate/fleet.
+- Do not publish or deploy as part of batching.
+
+## Workflow
+1. Partition into bounded units.
+2. Run units with existing agent/worktree tools.
+3. Integrate results and verify.

crates/tui/assets/skills/dataviz/SKILL.md

@@ -0,0 +1,20 @@
+---
+name: dataviz
+description: Choose and produce an appropriate chart/dashboard/visual explanation from data, with legible encodings and source/assumption notes.
+invocation: model+user
+---
+
+# Dataviz
+
+## When to use
+Use when the user needs a chart, dashboard, or visual explanation of data.
+
+## Non-goals
+- Do not choose chart types that hide uncertainty.
+- Do not invent data.
+
+## Workflow
+1. Inspect the data and questions.
+2. Choose encodings for the comparison at hand.
+3. Produce the visual with source/assumption notes.
+4. Sanity-check axes, units, and missing values.

crates/tui/assets/skills/debug/SKILL.md

@@ -0,0 +1,28 @@
+---
+name: debug
+description: Reproduce, minimize, localize, identify root cause, and distinguish diagnosis from an authorized fix. Prefer root-cause over symptom patches.
+invocation: model+user
+---
+
+# Debug
+
+## When to use
+Use for failures, flakes, wrong outputs, crashes, and regressions.
+
+## Non-goals
+- Do not jump to a fix before reproduction/localization when the bug is unclear.
+- Do not treat “run all tests” as debugging by default.
+
+## Workflow
+1. Reproduce or gather the strongest available failure signal.
+2. Minimize the case.
+3. Localize to component/file/line.
+4. State root cause vs symptoms.
+5. Fix only when authorized; otherwise hand back a diagnosis with evidence.
+
+## Output shape
+- Symptom
+- Reproduction
+- Localization
+- Root cause
+- Fix status / next step

crates/tui/assets/skills/dependency-update/SKILL.md

@@ -0,0 +1,20 @@
+---
+name: dependency-update
+description: Read release notes/changelogs, update a defined dependency scope, handle breaking changes, and verify. Explicit-only; no broad update-everything by inference.
+invocation: explicit-only
+---
+
+# Dependency Update
+
+## Invocation
+Explicit-only with a defined dependency scope.
+
+## Non-goals
+- No inferred “update everything”.
+- No publish/release.
+
+## Workflow
+1. Identify the dependency set and current versions.
+2. Read release notes/changelogs.
+3. Apply updates and fix breakages.
+4. Verify with the project’s test/build gates.

crates/tui/assets/skills/document/SKILL.md

@@ -0,0 +1,20 @@
+---
+name: document
+description: Write or update repository/product documentation: README, user guides, API docs, architecture, migration notes, and changelog material. This is not the Word-document artifact skill.
+invocation: model+user
+---
+
+# Document
+
+## When to use
+Use for repo/product docs: README, guides, API docs, architecture, migrations, changelog notes.
+
+## Non-goals
+- Not Word/DOCX authoring (use docx).
+- Not release publication (use release only when explicit).
+
+## Workflow
+1. Identify the doc audience and location.
+2. Prefer updating existing docs over creating duplicates.
+3. Keep commands and paths verified.
+4. Call out unknowns instead of inventing them.

crates/tui/assets/skills/documents/SKILL.md

@@ -1,31 +1,11 @@
 ---
 name: documents
-description: Create, edit, inspect, or convert Word documents and DOCX deliverables such as memos, reports, letters, templates, and forms.
+description: Compatibility alias for the docx skill. Prefer loading docx.
+aliases-for: docx
 ---
 
-# Documents
+# documents (alias)
 
-Use this skill when the user wants a `.docx` or Word-style document, or when an
-existing document needs edits, comments, extraction, or conversion.
+This skill is a compatibility alias for **docx**.
 
-## Workflow
-
-1. Clarify the output path if needed. Prefer a `.docx` file in the workspace.
-2. Preserve existing documents. Write edited copies instead of overwriting
-   originals unless the user asked for an in-place edit.
-3. Use the best available local tool:
-   - `python-docx` for creation and structured edits
-   - `pandoc` for Markdown/HTML to DOCX conversion
-   - OOXML inspection with `unzip` only for targeted low-level fixes
-4. For new documents, choose a simple professional structure: title, metadata
-   when useful, clear headings, readable body text, and tables only for real
-   row/column data.
-5. For edits, make the smallest change that satisfies the request and keep the
-   original formatting where practical.
-6. Verify by reopening the file with a parser such as `python-docx`, checking
-   paragraph/table counts, and extracting representative text. If LibreOffice
-   or another renderer is available, render or export a quick visual check.
-
-If required dependencies are missing, ask before installing packages. If the
-user only needs content and a DOCX cannot be produced in the environment, offer
-Markdown or HTML as a fallback and clearly state the limitation.
+Load and follow the `docx` skill body for the actual workflow.

crates/tui/assets/skills/docx/SKILL.md

@@ -0,0 +1,19 @@
+---
+name: docx
+description: Create/edit/inspect Word documents using available managed or host capabilities.
+invocation: model+user
+---
+
+# Docx
+
+## When to use
+Use for `.docx` / Word-style documents (memos, reports, letters, templates).
+
+## Compatibility
+`documents` is a compatibility alias for this skill.
+
+## Workflow
+1. Prefer `.docx` in the workspace.
+2. Preserve originals unless in-place edit is requested.
+3. Use python-docx/pandoc when available.
+4. Verify by reopening and extracting representative text.

crates/tui/assets/skills/frontend-design/SKILL.md

@@ -0,0 +1,20 @@
+---
+name: frontend-design
+description: Produce intentional, responsive, accessible UI work and perform visual QA instead of generic component assembly.
+invocation: model+user
+---
+
+# Frontend Design
+
+## When to use
+Use for UI that must feel intentional, responsive, and accessible.
+
+## Non-goals
+- Do not ship generic placeholder layouts as finished design.
+- Do not ignore accessibility.
+
+## Workflow
+1. Clarify audience and constraints.
+2. Design hierarchy, spacing, and states.
+3. Implement with responsive/accessible patterns.
+4. Visual QA against the acceptance criteria.

crates/tui/assets/skills/implement/SKILL.md

@@ -0,0 +1,24 @@
+---
+name: implement
+description: Carry an authorized, defined request or approved plan through scoped edits and proportionate verification. Skill loading is not write authorization by itself.
+invocation: model+user
+---
+
+# Implement
+
+## When to use
+Use for scoped implementation of a clear request or approved plan.
+
+## Non-goals
+- Loading this skill is not write authorization by itself.
+- Do not expand scope without saying so.
+- Do not skip verification after risky edits.
+
+## Workflow
+1. Confirm scope and acceptance criteria.
+2. Make the smallest coherent change set.
+3. Run the narrowest useful verification.
+4. Report what changed and what remains.
+
+## Failure mode
+If blocked by missing permissions or failing verification, stop and report evidence.

crates/tui/assets/skills/interview/SKILL.md

@@ -0,0 +1,24 @@
+---
+name: interview
+description: Ask one useful structured question at a time only when material product/implementation choices are genuinely missing; remember answers and produce a brief/spec. Discoverable facts should be investigated instead of asked.
+invocation: model+user
+---
+
+# Interview
+
+## When to use
+Use when the request lacks a material product or implementation choice that cannot be discovered from the workspace or tools.
+
+## Non-goals
+- Do not interview for facts you can read, search, or measure.
+- Do not ask multi-question dumps. One focused question at a time.
+- Do not invent requirements.
+
+## Workflow
+1. State the missing decision and why it blocks progress.
+2. Offer concrete options when possible.
+3. Record the answer and continue the work.
+4. Produce a short brief/spec only when enough answers exist.
+
+## Failure mode
+If the user declines to answer, proceed with explicit assumptions and label them.