diff --git a/.agents/skills/tiny-ps-chapter-test-loop/SKILL.md b/.agents/skills/tiny-ps-chapter-test-loop/SKILL.md
new file mode 100644
index 0000000..fa4b74d
--- /dev/null
+++ b/.agents/skills/tiny-ps-chapter-test-loop/SKILL.md
@@ -0,0 +1,20 @@
+---
+name: tiny-ps-chapter-test-loop
+description: Use when changing a Tiny PowerShell Projects chapter and you need the expected validation loop from chapter tests to full regression.
+---
+
+# Tiny PowerShell Chapter Test Loop
+
+## When to use
+- You are editing a chapter script or `solution*.ps1` file.
+- You want the fastest feedback loop before broader regression.
+
+## Workflow
+1. Work inside the chapter directory and keep `solution1.ps1` intact as a known baseline when possible.
+2. Run `./AllTest.ps1` in that chapter to execute `test.ps1` against each `solution*.ps1`.
+3. When the chapter passes, run `pwsh -NoProfile -File ./RunAllTests.ps1` from the repository root.
+4. Note unrelated existing failures separately instead of changing other chapters opportunistically.
+
+## Output
+- Chapter-local Pester results for all `solution*.ps1` variants.
+- A repository-wide regression result before finalizing changes.
diff --git a/.agents/skills/tiny-ps-deterministic-random/SKILL.md b/.agents/skills/tiny-ps-deterministic-random/SKILL.md
new file mode 100644
index 0000000..30e05a8
--- /dev/null
+++ b/.agents/skills/tiny-ps-deterministic-random/SKILL.md
@@ -0,0 +1,20 @@
+---
+name: tiny-ps-deterministic-random
+description: Use when a chapter relies on randomness and you need reproducible PowerShell behavior plus stable Pester assertions.
+---
+
+# Tiny PowerShell Deterministic Random
+
+## When to use
+- The script uses `Get-Random` or `[Random]`.
+- Tests or examples need exact output for a known seed.
+
+## Workflow
+1. Add a seed parameter and support the chapter's expected short and long flag names when tests or README require them.
+2. Initialize randomness exactly once near the start with `Get-Random -SetSeed` or `[Random]::new($seed)`.
+3. Keep the sequence of random calls stable so the same seed keeps producing the same output.
+4. Add Pester examples that assert exact results for chosen seeds and representative option combinations.
+
+## Output
+- Stable, repeatable output for the same seed.
+- Pester tests that lock in deterministic random behavior.
diff --git a/.agents/skills/tiny-ps-input-normalization/SKILL.md b/.agents/skills/tiny-ps-input-normalization/SKILL.md
new file mode 100644
index 0000000..457ed77
--- /dev/null
+++ b/.agents/skills/tiny-ps-input-normalization/SKILL.md
@@ -0,0 +1,20 @@
+---
+name: tiny-ps-input-normalization
+description: Use when a chapter accepts either literal text or a file path and should normalize input once before processing.
+---
+
+# Tiny PowerShell Input Normalization
+
+## When to use
+- The script can receive literal text or a path to a file.
+- The rest of the logic should not care where the content came from.
+
+## Workflow
+1. Accept a single text-or-path parameter unless the chapter already defines a clearer contract.
+2. Use `Test-Path` early to detect file input and read content once with `Get-Content -Raw` for whole-text transforms or line-based reads only when required.
+3. Normalize the input near the top so downstream logic works on content, not file paths.
+4. Add tests for both direct text and file input using the same expected output where appropriate.
+
+## Output
+- One normalized content flow regardless of whether input came from a file or the command line.
+- Paired tests covering both input forms.
diff --git a/.agents/skills/tiny-ps-regex-text-transform/SKILL.md b/.agents/skills/tiny-ps-regex-text-transform/SKILL.md
new file mode 100644
index 0000000..f32aead
--- /dev/null
+++ b/.agents/skills/tiny-ps-regex-text-transform/SKILL.md
@@ -0,0 +1,20 @@
+---
+name: tiny-ps-regex-text-transform
+description: Use when a chapter transforms text with regex while preserving delimiters, punctuation, and casing expectations.
+---
+
+# Tiny PowerShell Regex Text Transform
+
+## When to use
+- You are mutating words or characters inside larger text.
+- Whitespace-only splitting would lose punctuation or structure.
+
+## Workflow
+1. Identify whether the chapter needs token replacement (`[regex]::Replace`) or token-preserving splitting (`[regex]::Split`).
+2. Transform only the matched text, leaving separators and non-target characters untouched.
+3. Preserve casing intentionally when examples or tests expect it.
+4. Add examples for punctuation, apostrophes, short words, and mixed-case input.
+
+## Output
+- Text transforms that preserve surrounding structure.
+- Focused tests for punctuation and edge-case tokens.
diff --git a/.claude/commands/tiny-ps-chapter-test-loop.md b/.claude/commands/tiny-ps-chapter-test-loop.md
new file mode 100644
index 0000000..3171178
--- /dev/null
+++ b/.claude/commands/tiny-ps-chapter-test-loop.md
@@ -0,0 +1,21 @@
+---
+description: Use when changing a Tiny PowerShell Projects chapter and you need the expected validation loop from chapter tests to full regression.
+argument-hint: "[chapter directory]"
+allowed-tools: Read, Write, Edit, Glob, Grep, Bash(pwsh:*)
+---
+
+# Tiny PowerShell Chapter Test Loop
+
+## When to use
+- You are editing a chapter script or `solution*.ps1` file.
+- You want the fastest feedback loop before broader regression.
+
+## Workflow
+1. Work inside the chapter directory and keep `solution1.ps1` intact as a known baseline when possible.
+2. Run `./AllTest.ps1` in that chapter to execute `test.ps1` against each `solution*.ps1`.
+3. When the chapter passes, run `pwsh -NoProfile -File ./RunAllTests.ps1` from the repository root.
+4. Note unrelated existing failures separately instead of changing other chapters opportunistically.
+
+## Output
+- Chapter-local Pester results for all `solution*.ps1` variants.
+- A repository-wide regression result before finalizing changes.
diff --git a/.claude/commands/tiny-ps-deterministic-random.md b/.claude/commands/tiny-ps-deterministic-random.md
new file mode 100644
index 0000000..5dfdc24
--- /dev/null
+++ b/.claude/commands/tiny-ps-deterministic-random.md
@@ -0,0 +1,21 @@
+---
+description: Use when a chapter relies on randomness and you need reproducible PowerShell behavior plus stable Pester assertions.
+argument-hint: "[random behavior to make deterministic]"
+allowed-tools: Read, Write, Edit, Glob, Grep, Bash(pwsh:*)
+---
+
+# Tiny PowerShell Deterministic Random
+
+## When to use
+- The script uses `Get-Random` or `[Random]`.
+- Tests or examples need exact output for a known seed.
+
+## Workflow
+1. Add a seed parameter and support the chapter's expected short and long flag names when tests or README require them.
+2. Initialize randomness exactly once near the start with `Get-Random -SetSeed` or `[Random]::new($seed)`.
+3. Keep the sequence of random calls stable so the same seed keeps producing the same output.
+4. Add Pester examples that assert exact results for chosen seeds and representative option combinations.
+
+## Output
+- Stable, repeatable output for the same seed.
+- Pester tests that lock in deterministic random behavior.
diff --git a/.claude/commands/tiny-ps-input-normalization.md b/.claude/commands/tiny-ps-input-normalization.md
new file mode 100644
index 0000000..b58315d
--- /dev/null
+++ b/.claude/commands/tiny-ps-input-normalization.md
@@ -0,0 +1,21 @@
+---
+description: Use when a chapter accepts either literal text or a file path and should normalize input once before processing.
+argument-hint: "[input contract]"
+allowed-tools: Read, Write, Edit, Glob, Grep, Bash(pwsh:*)
+---
+
+# Tiny PowerShell Input Normalization
+
+## When to use
+- The script can receive literal text or a path to a file.
+- The rest of the logic should not care where the content came from.
+
+## Workflow
+1. Accept a single text-or-path parameter unless the chapter already defines a clearer contract.
+2. Use `Test-Path` early to detect file input and read content once with `Get-Content -Raw` for whole-text transforms or line-based reads only when required.
+3. Normalize the input near the top so downstream logic works on content, not file paths.
+4. Add tests for both direct text and file input using the same expected output where appropriate.
+
+## Output
+- One normalized content flow regardless of whether input came from a file or the command line.
+- Paired tests covering both input forms.
diff --git a/.claude/commands/tiny-ps-regex-text-transform.md b/.claude/commands/tiny-ps-regex-text-transform.md
new file mode 100644
index 0000000..2f18d62
--- /dev/null
+++ b/.claude/commands/tiny-ps-regex-text-transform.md
@@ -0,0 +1,21 @@
+---
+description: Use when a chapter transforms text with regex while preserving delimiters, punctuation, and casing expectations.
+argument-hint: "[text transform behavior]"
+allowed-tools: Read, Write, Edit, Glob, Grep, Bash(pwsh:*)
+---
+
+# Tiny PowerShell Regex Text Transform
+
+## When to use
+- You are mutating words or characters inside larger text.
+- Whitespace-only splitting would lose punctuation or structure.
+
+## Workflow
+1. Identify whether the chapter needs token replacement (`[regex]::Replace`) or token-preserving splitting (`[regex]::Split`).
+2. Transform only the matched text, leaving separators and non-target characters untouched.
+3. Preserve casing intentionally when examples or tests expect it.
+4. Add examples for punctuation, apostrophes, short words, and mixed-case input.
+
+## Output
+- Text transforms that preserve surrounding structure.
+- Focused tests for punctuation and edge-case tokens.
diff --git a/.github/skills/tiny-ps-chapter-test-loop/SKILL.md b/.github/skills/tiny-ps-chapter-test-loop/SKILL.md
new file mode 100644
index 0000000..fa4b74d
--- /dev/null
+++ b/.github/skills/tiny-ps-chapter-test-loop/SKILL.md
@@ -0,0 +1,20 @@
+---
+name: tiny-ps-chapter-test-loop
+description: Use when changing a Tiny PowerShell Projects chapter and you need the expected validation loop from chapter tests to full regression.
+---
+
+# Tiny PowerShell Chapter Test Loop
+
+## When to use
+- You are editing a chapter script or `solution*.ps1` file.
+- You want the fastest feedback loop before broader regression.
+
+## Workflow
+1. Work inside the chapter directory and keep `solution1.ps1` intact as a known baseline when possible.
+2. Run `./AllTest.ps1` in that chapter to execute `test.ps1` against each `solution*.ps1`.
+3. When the chapter passes, run `pwsh -NoProfile -File ./RunAllTests.ps1` from the repository root.
+4. Note unrelated existing failures separately instead of changing other chapters opportunistically.
+
+## Output
+- Chapter-local Pester results for all `solution*.ps1` variants.
+- A repository-wide regression result before finalizing changes.
diff --git a/.github/skills/tiny-ps-deterministic-random/SKILL.md b/.github/skills/tiny-ps-deterministic-random/SKILL.md
new file mode 100644
index 0000000..30e05a8
--- /dev/null
+++ b/.github/skills/tiny-ps-deterministic-random/SKILL.md
@@ -0,0 +1,20 @@
+---
+name: tiny-ps-deterministic-random
+description: Use when a chapter relies on randomness and you need reproducible PowerShell behavior plus stable Pester assertions.
+---
+
+# Tiny PowerShell Deterministic Random
+
+## When to use
+- The script uses `Get-Random` or `[Random]`.
+- Tests or examples need exact output for a known seed.
+
+## Workflow
+1. Add a seed parameter and support the chapter's expected short and long flag names when tests or README require them.
+2. Initialize randomness exactly once near the start with `Get-Random -SetSeed` or `[Random]::new($seed)`.
+3. Keep the sequence of random calls stable so the same seed keeps producing the same output.
+4. Add Pester examples that assert exact results for chosen seeds and representative option combinations.
+
+## Output
+- Stable, repeatable output for the same seed.
+- Pester tests that lock in deterministic random behavior.
diff --git a/.github/skills/tiny-ps-input-normalization/SKILL.md b/.github/skills/tiny-ps-input-normalization/SKILL.md
new file mode 100644
index 0000000..457ed77
--- /dev/null
+++ b/.github/skills/tiny-ps-input-normalization/SKILL.md
@@ -0,0 +1,20 @@
+---
+name: tiny-ps-input-normalization
+description: Use when a chapter accepts either literal text or a file path and should normalize input once before processing.
+---
+
+# Tiny PowerShell Input Normalization
+
+## When to use
+- The script can receive literal text or a path to a file.
+- The rest of the logic should not care where the content came from.
+
+## Workflow
+1. Accept a single text-or-path parameter unless the chapter already defines a clearer contract.
+2. Use `Test-Path` early to detect file input and read content once with `Get-Content -Raw` for whole-text transforms or line-based reads only when required.
+3. Normalize the input near the top so downstream logic works on content, not file paths.
+4. Add tests for both direct text and file input using the same expected output where appropriate.
+
+## Output
+- One normalized content flow regardless of whether input came from a file or the command line.
+- Paired tests covering both input forms.
diff --git a/.github/skills/tiny-ps-regex-text-transform/SKILL.md b/.github/skills/tiny-ps-regex-text-transform/SKILL.md
new file mode 100644
index 0000000..f32aead
--- /dev/null
+++ b/.github/skills/tiny-ps-regex-text-transform/SKILL.md
@@ -0,0 +1,20 @@
+---
+name: tiny-ps-regex-text-transform
+description: Use when a chapter transforms text with regex while preserving delimiters, punctuation, and casing expectations.
+---
+
+# Tiny PowerShell Regex Text Transform
+
+## When to use
+- You are mutating words or characters inside larger text.
+- Whitespace-only splitting would lose punctuation or structure.
+
+## Workflow
+1. Identify whether the chapter needs token replacement (`[regex]::Replace`) or token-preserving splitting (`[regex]::Split`).
+2. Transform only the matched text, leaving separators and non-target characters untouched.
+3. Preserve casing intentionally when examples or tests expect it.
+4. Add examples for punctuation, apostrophes, short words, and mixed-case input.
+
+## Output
+- Text transforms that preserve surrounding structure.
+- Focused tests for punctuation and edge-case tokens.
diff --git a/AI_WORKFLOW_SHORTLIST.md b/AI_WORKFLOW_SHORTLIST.md
new file mode 100644
index 0000000..8f0b6a5
--- /dev/null
+++ b/AI_WORKFLOW_SHORTLIST.md
@@ -0,0 +1,43 @@
+# AI Workflow Packaging Shortlist
+
+## Compact shortlist
+
+| Repeated workflow | Supporting evidence | Frequency / confidence | Recommended form | Why create (or skip) |
+| --- | --- | --- | --- | --- |
+| Run chapter-level tests while iterating on `solution*.ps1` | `AllTest.ps1` exists in 22 chapter folders; each chapter has `test.ps1`; root README explains the `solution*.ps1` loop | 22 chapters / High | Extend existing | Already automated well by chapter `AllTest.ps1`; no new asset needed |
+| Run full regression after local chapter changes | `RunAllTests.ps1` loops all numbered chapter folders and executes `AllTest.ps1`; README emphasizes running tests after each change | Repository-wide / High | Skill/playbook | Worth documenting as a consistent maintainer + agent workflow |
+| Author the same AI skill across multiple harness layouts | Open issues `#12`, `#14`, `#16`, `#18`; related open PRs `#13`, `#15`, `#17`, `#19`; active `copilot/*ai-skills*` branches | Repeated in recent repo activity / High | Script | Manual copy/edit across `.github/.agents/.claude` is repetitive and error-prone |
+| Add release/changelog automation | No release workflow, tags pipeline, or changelog process found in this branch | Low evidence / Low | Skip | Needs stronger maintainer signal before introducing publishing automation |
+
+## Created high-confidence missing items
+
+### `New-AISkillScaffold.ps1` (script)
+
+Creates a synchronized starter skill in all three harness locations from one command:
+
+- `.github/skills/<skill-name>/SKILL.md`
+- `.agents/skills/<skill-name>/SKILL.md`
+- `.claude/commands/<skill-name>.md`
+
+This packages the repeated cross-harness AI skill authoring workflow into one repeatable step with clear outputs.
+
+### Usage
+
+```powershell
+pwsh -NoProfile -File ./New-AISkillScaffold.ps1 \
+  -SkillName tiny-ps-example \
+  -Description "Use when implementing a repeatable Tiny PowerShell Projects workflow."
+```
+
+Use `-Force` to overwrite existing scaffolded files.
+
+### Repository-derived skill set
+
+- `tiny-ps-chapter-test-loop`
+  - Packages the repeated chapter-first, repo-second validation workflow documented in `README.md` and implemented by `AllTest.ps1` plus `RunAllTests.ps1`.
+- `tiny-ps-deterministic-random`
+  - Encodes the repeated seeded-random pattern used in chapters such as `09_abuse`, `10_telephone`, `12_ransom`, `16_scrambler`, `19_wod`, and `20_password`.
+- `tiny-ps-regex-text-transform`
+  - Captures the repeated regex-first text transformation approach used in `08_apples_and_bananas`, `15_kentucky_friar`, `16_scrambler`, and `17_mad_libs`.
+- `tiny-ps-input-normalization`
+  - Packages the recurring "literal text or file path" input handling found in chapters such as `05_howler`, `10_telephone`, `12_ransom`, and `15_kentucky_friar`.
diff --git a/New-AISkillScaffold.ps1 b/New-AISkillScaffold.ps1
new file mode 100644
index 0000000..d2a97b3
--- /dev/null
+++ b/New-AISkillScaffold.ps1
@@ -0,0 +1,97 @@
+param(
+    [Parameter(Mandatory)]
+    [ValidatePattern('^[a-z0-9][a-z0-9-]*$')]
+    [string]$SkillName,
+
+    [Parameter(Mandatory)]
+    [string]$Description,
+
+    [string]$Title,
+
+    [string]$ArgumentHint = '[task context]',
+
+    [switch]$Force
+)
+
+Set-StrictMode -Version Latest
+$ErrorActionPreference = 'Stop'
+
+if ([string]::IsNullOrWhiteSpace($Title)) {
+    $Title = ((($SkillName -replace '-', ' ').Split(' ', [System.StringSplitOptions]::RemoveEmptyEntries) | ForEach-Object {
+                if ($_.Length -gt 1) { $_.Substring(0, 1).ToUpper() + $_.Substring(1) }
+                else { $_.ToUpper() }
+            }) -join ' ')
+}
+
+function Write-TemplateFile {
+    param(
+        [Parameter(Mandatory)]
+        [string]$Path,
+
+        [Parameter(Mandatory)]
+        [string]$Content
+    )
+
+    $directory = Split-Path -Path $Path -Parent
+    if ($directory) {
+        New-Item -Path $directory -ItemType Directory -Force | Out-Null
+    }
+
+    if ((Test-Path -Path $Path) -and -not $Force) {
+        throw "File already exists: $Path`nUse -Force to overwrite."
+    }
+
+    Set-Content -Path $Path -Value $Content
+}
+
+$sharedSkillBody = @"
+---
+name: $SkillName
+description: $Description
+---
+
+# $Title
+
+## When to use
+- TODO: Describe trigger conditions.
+
+## Workflow
+1. TODO: Define stable inputs.
+2. TODO: Execute repeatable steps.
+3. TODO: Validate clear output or stopping condition.
+
+## Output
+- TODO: Describe expected output.
+"@
+
+$claudeBody = @"
+---
+description: $Description
+argument-hint: $ArgumentHint
+allowed-tools: Read, Write, Edit, Glob, Grep, Bash(pwsh:*)
+---
+
+# $Title
+
+## When to use
+- TODO: Describe trigger conditions.
+
+## Workflow
+1. TODO: Define stable inputs.
+2. TODO: Execute repeatable steps.
+3. TODO: Validate clear output or stopping condition.
+
+## Output
+- TODO: Describe expected output.
+"@
+
+$targets = @(
+    @{ Path = ".github/skills/$SkillName/SKILL.md"; Content = $sharedSkillBody },
+    @{ Path = ".agents/skills/$SkillName/SKILL.md"; Content = $sharedSkillBody },
+    @{ Path = ".claude/commands/$SkillName.md"; Content = $claudeBody }
+)
+
+foreach ($target in $targets) {
+    Write-TemplateFile -Path $target.Path -Content $target.Content
+    Write-Host "Created $($target.Path)"
+}