output-style-builder

Original source / Raw SKILL.md

---
name: output-style-builder
type: standard
depth: extended
description: >-
  Creates structured output formats (JSON, YAML, Markdown-KV, XML) and response style
  configurations for Claude. Use when: (1) defining agent output schemas (.json, .yaml, .xml),
  (2) configuring response style scope hierarchy, (3) embedding formats in agents/commands,
  (4) building CLAUDE.md output sections, (5) creating format definitions in .claude/styles/,
  or (6) standardizing structured data serialization.
---

# [H1][OUTPUT-STYLE-BUILDER]
>**Dictum:** *Parsing clarity requires structured serialization.*

<br>

Build agent response formats. Configure scope-based style.

[DELEGATE] Voice, formatting, constraint rules → `style-standards` skill.

**Tasks:**
1. Read [index.md](./index.md) — Reference file listing for navigation
2. Select domain — FORMATS (structured data) or CONFIGURATION (response style scope)
3. (formats) Read [formats.md](./references/formats.md) — Selection metrics, embedding, validation scoring
4. (configuration) Read [configuration.md](./references/configuration.md) — Scope hierarchy, embedding patterns
5. (schema) Read [schema.md](./references/schema.md) — Delimiter syntax, canonical examples
6. (structure) Read [structure.md](./references/structure.md) — Ordering patterns, composition
7. (prose) Load `style-standards` skill — Voice, formatting, constraints
8. Execute per domain — Apply Guidance and Best-Practices
9. Validate — Quality gate; see §VALIDATION

**Scope:**
- *Formats:* Structured data serialization (JSON, YAML, Markdown-KV, XML) for agent output.
- *Configuration:* Response style scope hierarchy (global → project → skill → command).

**Domain Navigation:**
- *[FORMATS]* — Structured data output. Load for: agent schemas, API responses, validation scoring.
- *[CONFIGURATION]* — Scope hierarchy. Load for: CLAUDE.md output sections, precedence rules.
- *[SCHEMA]* — Delimiters, examples. Load for: syntax reference, canonical patterns.
- *[STRUCTURE]* — Ordering, composition. Load for: section sequencing, chaining.

[REFERENCE]: [index.md](./index.md) — Complete reference file listing

---
## [1][FORMATS]
>**Dictum:** *Format optimization requires accuracy-token tradeoff analysis.*

<br>

[IMPORTANT] Format choice impacts accuracy (16pp variance) and tokens (6.75x variance).

**Guidance:**
- `Selection` — Markdown-KV: 60.7% accuracy, 2.7x tokens. JSON: 52.3% accuracy, 0.85x tokens.
- `Embedding` — Inline for single use. Reference (`@.claude/styles/`) for 3+ consumers.
- `Validation` — 100-point scoring. Deployment requires score >= 80.

**Best-Practices:**<br>
- Single format per output type.
- Max 5 variables per format; every optional has default.
- Constrained decoding guarantees 97-100% schema compliance.

**References:**<br>
- [→formats.md](./references/formats.md): Selection, embedding, validation.
- [→format.template.md](./templates/format.template.md): Format scaffold.

---
## [2][CONFIGURATION]
>**Dictum:** *Narrow context requires specialized output rules.*

<br>

[IMPORTANT] Higher precedence overrides lower (command level 5 > global level 1).

**Guidance:**
- `Global` — CLAUDE.md `[OUTPUT]` section. 50-100 LOC optimal.
- `Project` — PROJECT.md overrides. Document divergence reason.
- `Skill/Agent` — Inline or reference pattern. Match reuse requirements.
- `Command` — Narrowest scope. Specialized output for single invocation.

**Best-Practices:**<br>
- Global rules in CLAUDE.md; overrides minimal (max 30 LOC).
- Reference `style-standards` for voice/formatting rules—no duplication.
- Place weight-10 constraints first in all scopes.

**References:**<br>
- [→configuration.md](./references/configuration.md): Scope hierarchy, embedding.
- [→style.template.md](./templates/style.template.md): Style scaffold.
- [DELEGATE] `style-standards` skill: Voice, formatting, constraints.

---
## [3][SCHEMA]
>**Dictum:** *Performance consistency requires delimiter standardization.*

<br>

[IMPORTANT] Maintain delimiter consistency throughout output.

**Guidance:**
- `Delimiters` — Code fence (` ``` `), separator (`---`), soft break (`<br>`).
- `Consistency` — Prohibit mixed delimiter styles within single output.

**References:**<br>
- [→schema.md](./references/schema.md): Syntax reference.

---
## [4][STRUCTURE]
>**Dictum:** *Attention decay requires priority-first placement.*

<br>

[IMPORTANT] Primacy effect assigns 5.79x weight to early items.

**Guidance:**
- `Ordering` — Action-first, Priority-first, or Context-first patterns.
- `Hierarchy` — Maximum 3 levels. 2-7 items per container.
- `Composition` — Base-override inheritance. Shallow or deep merge.

**References:**<br>
- [→structure.md](./references/structure.md): Sequencing, composability.

---
## [5][TEMPLATES]
>**Dictum:** *Structural consistency requires scaffold reuse.*

<br>

**FORMATS Domain:**<br>
- [→format.template.md](./templates/format.template.md): Structured data format scaffold.

**CONFIGURATION Domain:**<br>
- [→style.template.md](./templates/style.template.md): Response style scaffold.

---
## [6][VALIDATION]
>**Dictum:** *Gates prevent incomplete artifacts.*

<br>

[VERIFY] Completion:
- [ ] Domain: Selected FORMATS or CONFIGURATION scope.
- [ ] References: Required domain files loaded per Tasks.
- [ ] Style: `style-standards` delegation applied—zero duplication.
- [ ] Format: Validation score >= 80 (formats domain).
- [ ] Quality: LOC within limits, weight-10 constraints first.

[REFERENCE] Operational checklist: [→validation.md](./references/validation.md)


---

## Referenced Files

> The following files are referenced in this skill and included for context.

### index.md

```markdown
# [H1][INDEX]
>**Dictum:** *Reference navigation requires centralized discovery.*

<br>

| [INDEX] | [DOMAIN]      | [PATH]                                              | [DICTUM]                                              |
| :-----: | ------------- | --------------------------------------------------- | ----------------------------------------------------- |
|   [1]   | Formats       | [→formats.md](references/formats.md)                | Accuracy-token tradeoffs require format governance.   |
|   [2]   | Configuration | [→configuration.md](references/configuration.md)    | Response style precedence follows scope hierarchy.    |
|   [3]   | Schema        | [→schema.md](references/schema.md)                  | Format parsing requires delimiter syntax anchors.     |
|   [4]   | Structure     | [→structure.md](references/structure.md)            | Format reuse requires ordering standards.             |
|   [5]   | Format TPL    | [→format.template.md](templates/format.template.md) | Structured data formats require scaffold templates.   |
|   [6]   | Style TPL     | [→style.template.md](templates/style.template.md)   | Response style configuration requires scaffold reuse. |
|   [7]   | Validation    | [→validation.md](references/validation.md)          | Operational criteria verify output quality.           |

[DELEGATE] Voice, formatting, constraints → `style-standards` skill.

```

### references/formats.md

```markdown
# [H1][FORMATS]
>**Dictum:** *Empirical format selection requires accuracy-token tradeoff analysis.*

<br>

[IMPORTANT] Format selection impacts accuracy by 16pp and token cost by 6.75x. Constrained decoding achieves 97-100% compliance.

---
## [1][SELECTION]
>**Dictum:** *Data-driven format choice requires 16pp accuracy awareness.*

<br>

| [INDEX] | [FORMAT]    | [ACCURACY] | [TOKEN_COST] | [STRENGTH]        | [USE_CASE]             |
| :-----: | ----------- | :--------: | :----------: | ----------------- | ---------------------- |
|   [1]   | Markdown-KV |   60.7%    |     2.7x     | Highest accuracy  | Agent output, reports  |
|   [2]   | XML         |   56.0%    |     1.8x     | Schema validation | Compliance, strict     |
|   [3]   | YAML        |   54.7%    |     0.7x     | Human-readable    | Config files           |
|   [4]   | JSON        |   52.3%    |    0.85x     | Universal parsing | API output, structured |
|   [5]   | CSV         |   44.3%    |     1.0x     | Minimal overhead  | Tabular data           |

**Token Cost Baseline:** CSV = 1.0x.

<br>

### [1.1][DECISION_DISPATCH]

```
1. Count consumers: 1 use → inline; 3+ uses → reference pattern.
2. Identify structure: flat → CSV; nested → JSON/YAML; strict schema → XML.
3. Accuracy need: critical → Markdown-KV; balanced → JSON.
4. Token budget: constrained → YAML (0.7x); flexible → Markdown-KV (2.7x).
```

---
### [1.2][SCHEMA_ENFORCEMENT]

| [INDEX] | [PROVIDER] | [METHOD]               | [COMPLIANCE] |
| :-----: | ---------- | ---------------------- | :----------: |
|   [1]   | OpenAI     | Constrained decoding   |     100%     |
|   [2]   | Anthropic  | Tool-based constraints |     100%     |
|   [3]   | XGrammar   | PDA-based (5x speedup) |   97-100%    |
|   [4]   | Outlines   | FSM-based              |    95-97%    |

[CRITICAL] Structured outputs guarantee format compliance, NOT semantic correctness.

---
## [2][WEIGHTING]
>**Dictum:** *5.79x attention differential requires priority-first placement.*

<br>

| [INDEX] | [WEIGHT] | [SEVERITY] | [ATTENTION] |
| :-----: | :------: | ---------- | :---------: |
|   [1]   |    10    | CRITICAL   |    5.79x    |
|   [2]   |   7-9    | IMPORTANT  |    3.2x     |
|   [3]   |   4-6    | STANDARD   |    1.0x     |
|   [4]   |   1-3    | OPTIONAL   |    0.5x     |

[IMPORTANT] Position weight-10 sections first. Place critical constraints at sequence start.

---
## [3][EMBEDDING]
>**Dictum:** *Duplication prevention across 3+ consumers requires reference pattern.*

<br>

| [INDEX] | [PATTERN]   | [SYNTAX]                    | [USE_CASE]              |
| :-----: | ----------- | --------------------------- | ----------------------- |
|   [1]   | Inline      | Embed full spec in body     | Single-use formats      |
|   [2]   | Reference   | `@.claude/styles/{name}.md` | Shared across 3+ agents |
|   [3]   | Conditional | Dispatch table by type      | Multi-format output     |

<br>

### [3.1][AGENT_EMBEDDING]

**Location:** Body preamble (after frontmatter, before first H2).

```markdown
---
name: agent-name
tools: Read, ...
---

[Purpose statement]

**Output Format:** @.claude/styles/{format-name}.md

---
## [1][PROCESS]
```

[IMPORTANT]:
- Include `Read` in agent `tools:` frontmatter for reference pattern.
- Embed format reference in body preamble. Exclude from H2 sections.

[CRITICAL] Reference pattern embeds entire file at `@` marker. Use inline embedding for single-use formats.

---
### [3.2][VARIABLES]

**Syntax:**<br>
- Required: `${variable-name}`
- Optional: `${variable-name:-default}`
- Conditional: `${variable-name?}`

**Resolution:** Compiler resolves frontmatter variables. Runtime resolves context values.

[REFERENCE] Validation checklist: [→validation.md§1](./validation.md#1formats_gate)

```

### references/configuration.md

```markdown
# [H1][CONFIGURATION]
>**Dictum:** *Hierarchical scope governance ensures response style consistency.*

<br>

[IMPORTANT] Configuration hierarchy: global (1) → project (2) → skill (3) → agent (4) → command (5). Higher precedence overrides lower.

[CRITICAL] **DELEGATE** Voice, formatting, constraint rules → `style-standards` skill. Scope hierarchy and embedding patterns covered here.

---
## [1][HIERARCHY]
>**Dictum:** *Five-level scope hierarchy enables granular control.*

<br>

| [INDEX] | [SCOPE] | [LOCATION]                | [PRECEDENCE] | [USE_CASE]           |
| :-----: | ------- | ------------------------- | :----------: | -------------------- |
|   [1]   | Global  | CLAUDE.md                 |      1       | All Claude responses |
|   [2]   | Project | PROJECT.md                |      2       | Project override     |
|   [3]   | Skill   | .claude/skills/*/SKILL.md |      3       | Skill-specific       |
|   [4]   | Agent   | .claude/agents/*.md       |      4       | Agent-specific       |
|   [5]   | Command | .claude/commands/*.md     |      5       | Command-specific     |

**Inheritance:** Base + override merge. Higher precedence wins on conflict.

---
## [2][GLOBAL]
>**Dictum:** *Global defaults require single source configuration.*

<br>

**Location:** `/CLAUDE.md` → `## [N][OUTPUT]` section.

**Constraints:**<br>
- 50-100 LOC for output section.
- Position weight-10 constraints first.
- Reference external files for content exceeding 100 LOC.

**Pattern:**
```markdown
## [4][OUTPUT]
>**Dictum:** *Output format optimizes readability.*

<br>

[IMPORTANT]:
- [ALWAYS] Use `backticks` for file paths, symbols, CLI commands.
- [ALWAYS] Avoid large code blocks—reference file/symbol names.
- [ALWAYS] Markdown: headings for structure, bullets for lists.

[CRITICAL]:
- [NEVER] Use emojis—use `[X]` style markers.
- [NEVER] Meta-commentary: "Sourced from...", "Confirmed with...".
```

[REFERENCE] Voice rules → `style-standards/references/voice/grammar.md`.

---
## [3][PROJECT]
>**Dictum:** *Specialized contexts require project-level overrides.*

<br>

**Location:** `PROJECT.md` or project-specific `.claude/` directories.

**Constraints:**<br>
- Inherit from global CLAUDE.md.
- Document divergence reason explicitly.
- Maximum 30 LOC for override section.

**Pattern:**
```markdown
## [OUTPUT]
>**Dictum:** *API output requires JSON format.*

<br>

[IMPORTANT] Override: All API responses use minified JSON.
- Rationale: Integration partners require machine-parseable output.
```

---
## [4][SKILL_AGENT]
>**Dictum:** *Skill/agent scope isolation enables task-specific output.*

<br>

**Location:** SKILL.md body preamble or agent frontmatter + body.

**Embedding Patterns:**<br>
- *Inline:* Embed full spec in body (single-use).
- *Reference:* `@.claude/styles/{name}.md` (3+ consumers).

**Reference Pattern:**
```markdown
---
name: agent-name
tools: Read, ...
---

[Purpose statement]

**Output Format:** @.claude/styles/{format-name}.md

---
## [1][PROCESS]
```

[IMPORTANT]:
- Include `Read` in agent `tools:` frontmatter for reference pattern.
- Embed format reference in body preamble. Exclude from H2 sections.

---
## [5][COMMAND]
>**Dictum:** *Narrowest scope enables single invocation specialization.*

<br>

**Location:** Command `## [N][AGENT_PROMPT_TEMPLATE]` section.

**Pattern:**
```markdown
## [5][AGENT_PROMPT_TEMPLATE]

@.claude/styles/report.md

[Command-specific instructions...]
```

**Use Case:** Specialized output for single command invocation (reports, API responses, templates).

---
## [6][STORAGE]
>**Dictum:** *Distinct storage directories enable format-style separation.*

<br>

| [INDEX] | [DIRECTORY]              | [SCOPE]     | [CONTENT]                  |
| :-----: | ------------------------ | ----------- | -------------------------- |
|   [1]   | `.claude/styles/`        | Agent-level | Data format definitions    |
|   [2]   | `.claude/output-styles/` | Global      | Response style definitions |

**`.claude/styles/`** — Structured data formats (JSON schema, Markdown-KV, YAML). Embedded via `@` in agent/command body.<br>
**`.claude/output-styles/`** — Prose response styles (voice, tone, structure). Referenced from CLAUDE.md or standalone.

[REFERENCE] Validation checklist: [→validation.md§2](./validation.md#2configuration_gate)

```

### references/schema.md

```markdown
# [H1][SCHEMA]
>**Dictum:** *Parsing reliability requires consistent delimiter syntax.*

<br>

[REFERENCE] Format selection and metrics → [→formats.md§1](./formats.md#1selection).

---
## [1][DELIMITERS]
>**Dictum:** *Delimiter consistency prevents 18-29% performance variance.*

<br>

[IMPORTANT] Inconsistent delimiters cause 18-29% performance variance.

| [INDEX] | [DELIMITER]  | [SYNTAX]          | [PURPOSE]                          |
| :-----: | ------------ | ----------------- | ---------------------------------- |
|   [1]   | Code fence   | ` ``` ` + lang    | Block boundary, language hint      |
|   [2]   | Separator    | `---`             | Section boundary (H2 to H2)        |
|   [3]   | Soft break   | `<br>`            | Transition (after Dictum/Preamble) |
|   [4]   | Inline code  | `` ` ``           | Symbol/path boundary               |
|   [5]   | Boundary tag | `[START]`/`[END]` | Explicit parse anchors             |

[CRITICAL] Prohibit mixed delimiter styles within single output.

---
## [2][EXAMPLES]
>**Dictum:** *Canonical reference patterns standardize format implementation.*

<br>

**JSON:**
```json
{
  "status": "success",
  "data": { "key": "value" },
  "errors": []
}
```

**Markdown-KV:**
```markdown
Status: success
Data:
  - key: value
Errors: none
```

**YAML:**
```yaml
status: success
data:
  key: value
errors: []
```

```

### references/structure.md

```markdown
# [H1][STRUCTURE]
>**Dictum:** *Compositional structure enables format reuse across contexts.*

<br>

[REFERENCE] Attention weights and ordering algorithm → [→formats.md§2](./formats.md#2weighting).

---
## [1][ORDERING]
>**Dictum:** *Impact-weighted ordering minimizes cognitive load.*

<br>

| [INDEX] | [PATTERN]      | [SEQUENCE]                                    | [USE_CASE]             |
| :-----: | -------------- | --------------------------------------------- | ---------------------- |
|   [1]   | Action-first   | Summary -> Blockers -> Details -> Context     | Immediate execution    |
|   [2]   | Priority-first | Failures -> Warnings -> Confirmations -> Info | Attention optimization |
|   [3]   | Context-first  | Scope -> Findings -> Details -> Action        | Understanding focus    |

---
## [2][HIERARCHY]
>**Dictum:** *Depth constraints prevent cognitive overload.*

<br>

[IMPORTANT] Limit hierarchy to 3 levels. Prohibit H4+.

| [INDEX] | [LEVEL] | [PURPOSE]           | [CONSTRAINT]           |
| :-----: | :-----: | ------------------- | ---------------------- |
|   [1]   |   L1    | Essential task info | Always visible         |
|   [2]   |   L2    | Supporting details  | Expandable/conditional |
|   [3]   |   L3    | Reference/optional  | Hidden by default      |

**Section Limits:**<br>
- Limit containers to 2-7 items.
- Limit levels to 3-5 constraints.
- Limit files to 10 markers maximum.

---
## [3][COMPOSITION]
>**Dictum:** *Inheritance patterns enable format reuse across outputs.*

<br>

**Base-Override Pattern:**
```yaml
base: ${base-style-name}
override:
  format: json
  sections:
    - name: status
      required: true
```

**Placeholder Syntax:**<br>
- Required: `${variable-name}`
- Optional: `${variable-name:-default}`
- Conditional: `${variable-name?}`

**Merge Semantics:**<br>
- Shallow: Override replaces entirely.
- Deep: Arrays concatenate, objects merge.
- Delete: `${property: null}` removes inherited.

---
## [4][CHAINING]
>**Dictum:** *Explicit stage coupling ensures pipeline reliability.*

<br>

| [INDEX] | [STAGE]   | [INPUT]         | [OUTPUT]        | [EXAMPLE]            |
| :-----: | --------- | --------------- | --------------- | -------------------- |
|   [1]   | Extract   | Raw response    | Structured data | Parse JSON from MD   |
|   [2]   | Transform | Structured data | Reformatted     | Convert JSON to YAML |
|   [3]   | Validate  | Reformatted     | Validated       | Schema compliance    |
|   [4]   | Emit      | Validated       | Final response  | Apply template       |

[CRITICAL] Design each stage to produce complete, valid output.

```

### templates/format.template.md

```markdown
---
name: ${format-name}
description: ${format-purpose}
format: ${json|yaml|xml|markdown-kv}
base: ${base-format:-none}
---

# [H1][${FORMAT_NAME}]
>**Dictum:** *${format-truth}.*

<br>

${one-sentence-purpose}

---
## [1][FORMAT]
>**Dictum:** *Consistent delimiter boundaries enable structured parsing.*

<br>

**Type:** `${format-type}`<br>
**Fence:** ` ``` ` + `${language-hint}`<br>
**Base:** `${base-format:-none}`<br>
**Schema:** `${schema-enforcement:-none}`

```${format-type}
${format-template}
```

---
## [2][SECTIONS]
>**Dictum:** *5.79x attention differential requires priority-first ordering.*

<br>

| [INDEX] | [SECTION]    | [REQUIRED] | [WEIGHT] | [PURPOSE]            |
| :-----: | ------------ | :--------: | :------: | -------------------- |
|   [1]   | ${section-1} |   ${y/n}   |    10    | ${section-1-purpose} |
|   [2]   | ${section-2} |   ${y/n}   |    8     | ${section-2-purpose} |
|   [3]   | ${section-3} |   ${y/n}   |    6     | ${section-3-purpose} |

[IMPORTANT]:
- [ALWAYS] Weight 10 = anchor position (58% attention).
- [ALWAYS] Sort descending by weight.
- [ALWAYS] Group equal-weight by logical flow.

---
## [3][VARIABLES]
>**Dictum:** *Dynamic value resolution enables runtime context.*

<br>

[IMPORTANT] Compiler resolves frontmatter variables. Runtime resolves context values.

| [INDEX] | [VARIABLE]    | [REQUIRED] | [DEFAULT]    | [DESCRIPTION]        |
| :-----: | ------------- | :--------: | ------------ | -------------------- |
|   [1]   | ${var-1-name} |     y      | none         | ${var-1-description} |
|   [2]   | ${var-2-name} |     n      | ${default-2} | ${var-2-description} |

**Syntax:**<br>
- Required: `${variable-name}`
- Optional: `${variable-name:-default}`
- Conditional: `${variable-name?}`

---
## [4][VALIDATION]
>**Dictum:** *Malformed output prevention requires quality gates.*

<br>

| [INDEX] | [DIMENSION]       | [POINTS] | [VALIDATOR]                          |
| :-----: | ----------------- | :------: | ------------------------------------ |
|   [1]   | Format valid      |    30    | Parser accepts (JSON/YAML/XML/MD-KV) |
|   [2]   | Delimiters        |    20    | Fence + separator consistency        |
|   [3]   | Required sections |    25    | All `required: y` sections present   |
|   [4]   | Variables         |    15    | All required variables resolved      |
|   [5]   | Anti-bloat        |    10    | No prohibited patterns               |
|         | **TOTAL**         | **100**  | Score >= 80 = pass                   |

<br>

### [4.1][ANTI_BLOAT]

**Prohibited Patterns:**
```regex
/Sourced from|Confirmed with|Based on/i  → meta-commentary
/This file|We do|You should/i            → self-reference
/might|could|probably|perhaps/i          → hedging
/\b(please|kindly|just|really)\b/i       → filler stopwords
```

[IMPORTANT] Prohibited patterns deduct 2 points per violation (max deduction: 10 points).

---
### [4.2][GATE]

[VERIFY]:
- [ ] Format: Output matches `${format-type}` specification.
- [ ] Delimiters: Fence syntax and separators consistent.
- [ ] Sections: All required sections present in correct order.
- [ ] Variables: All required variables resolved.
- [ ] Weights: Sections ordered descending by weight (10 → 1).
- [ ] Anti-bloat: No prohibited patterns detected.
- [ ] Score: Validation total >= 80 points.

```

### templates/style.template.md

```markdown
---
name: ${style-name}
description: ${style-purpose}
scope: ${global|project|skill|command}
audience: ${target-audience}
---

# [H1][${STYLE_NAME}]
>**Dictum:** *${response-truth}.*

<br>

${one-sentence-purpose}

[DELEGATE] Voice rules → `style-standards/references/voice/`. Formatting rules → `style-standards/references/formatting/`.

---
## [1][VOICE]
>**Dictum:** *Parameterized voice ensures consistent tone across contexts.*

<br>

| [INDEX] | [PARAMETER]     | [LEVEL]           | [EFFECT]                       |
| :-----: | --------------- | ----------------- | ------------------------------ |
|   [1]   | Formality       | ${0-100}%         | ${formality-description}       |
|   [2]   | Technical Depth | ${low\|med\|high} | ${technical-depth-description} |
|   [3]   | Directness      | ${0-100}%         | ${directness-description}      |
|   [4]   | Enthusiasm      | ${0-100}%         | ${enthusiasm-description}      |

<br>

### [1.1][GRAMMAR]

[IMPORTANT]:
- [ALWAYS] **Active Voice:** ${active-voice-requirement}.
- [ALWAYS] **Imperative Actions:** ${imperative-pattern}.
- [ALWAYS] **Tense:** ${tense-requirement}.

[CRITICAL]:
- [NEVER] **Self-Reference:** Prohibit "I", "we", "you" unless ${exception-condition}.
- [NEVER] **Hedging:** Prohibit "might", "could", "probably", "should".
- [NEVER] **Meta-Commentary:** Prohibit "Sourced from", "Confirmed with".

[REFERENCE] Grammar → `style-standards/references/voice/grammar.md`.

<br>

### [1.2][TONE]

**Primary:** ${primary-tone} — ${tone-description}.<br>
**Secondary:** ${secondary-tone} — ${tone-context}.<br>
**Prohibited:** ${prohibited-tone-1}, ${prohibited-tone-2}.

---
## [2][STRUCTURE]
>**Dictum:** *Structured formatting reduces cognitive load.*

<br>

| [INDEX] | [ELEMENT]        | [THRESHOLD]   | [RULE]                           |
| :-----: | ---------------- | ------------- | -------------------------------- |
|   [1]   | Response Length  | ${min}-${max} | ${length-constraint-description} |
|   [2]   | Paragraph Length | ${min}-${max} | ${paragraph-constraint}          |
|   [3]   | Header Usage     | ${threshold}  | ${header-usage-rule}             |
|   [4]   | List Conversion  | ${threshold}  | ${list-threshold-description}    |

[REFERENCE] Structure → `style-standards/references/formatting/structure.md`.

---
## [3][CONSTRAINTS]
>**Dictum:** *Explicit boundaries enforce output quality.*

<br>

| [INDEX] | [CATEGORY]   | [CONSTRAINT]          | [ENFORCEMENT]                     |
| :-----: | ------------ | --------------------- | --------------------------------- |
|   [1]   | Token Budget | ${min}-${max} tokens  | ${budget-enforcement-description} |
|   [2]   | Emojis       | ${allowed\|forbidden} | ${emoji-policy}                   |
|   [3]   | Code Blocks  | ${threshold}          | ${code-block-rule}                |

<br>

### [3.1][PROHIBITED]

[CRITICAL]:
- [NEVER] **${prohibited-pattern-1}:** ${rationale-1}.
- [NEVER] **${prohibited-pattern-2}:** ${rationale-2}.
- [NEVER] **${prohibited-pattern-3}:** ${rationale-3}.

[REFERENCE] Constraints → `style-standards/references/voice/constraints.md`.

---
## [4][VALIDATION]
>**Dictum:** *Pre-flight verification prevents deployment errors.*

<br>

[VERIFY]:
- [ ] Voice: Formality at ${formality}%, active voice, zero hedging.
- [ ] Structure: Response within ${min}-${max} words, headers per threshold.
- [ ] Constraints: No emojis, no meta-commentary, no self-reference.
- [ ] Tone: ${primary-tone} maintained throughout.

```

### references/validation.md

```markdown
# [H1][VALIDATION]
>**Dictum:** *Operational criteria verify output quality.*

<br>

Consolidated checklist for output-style-builder. SKILL.md §VALIDATION contains high-level gates; this file contains operational verification procedures.

---
## [1][FORMATS_GATE]
>**Dictum:** *Format deployment requires 80-point minimum.*

<br>

[VERIFY] Format compliance:
- [ ] Format selected matches use case strength (accuracy vs token tradeoff).
- [ ] Sections weighted and ordered by severity (10 → 1 descending).
- [ ] Embedding pattern matches reuse requirements (inline vs reference).
- [ ] Delimiter consistency verified throughout output.
- [ ] Required variables resolved; optionals have defaults.
- [ ] Validation score >= 80.

---
## [2][CONFIGURATION_GATE]
>**Dictum:** *Scope hierarchy prevents configuration conflicts.*

<br>

[VERIFY] Configuration compliance:
- [ ] Scope level matches use case (global → command hierarchy).
- [ ] CLAUDE.md output section <= 100 LOC.
- [ ] Override sections document divergence reason.
- [ ] Reference pattern uses `@` syntax with valid path.
- [ ] Reference files exist in `.claude/styles/` or `.claude/output-styles/`.
- [ ] Voice/formatting rules delegate to `style-standards`—zero duplication.
- [ ] Inheritance hierarchy respected (higher precedence wins).

---
## [3][SCORING]
>**Dictum:** *Quality gates prevent formats below 80% compliance.*

<br>

| [INDEX] | [DIMENSION]       | [POINTS] | [VALIDATOR]                          |
| :-----: | ----------------- | :------: | ------------------------------------ |
|   [1]   | Format valid      |    30    | Parser accepts (JSON/YAML/XML/MD-KV) |
|   [2]   | Required sections |    25    | All `required: true` present         |
|   [3]   | Delimiters        |    20    | Fence + separator consistency        |
|   [4]   | Variables         |    15    | All required vars resolved           |
|   [5]   | Anti-bloat        |    10    | No prohibited patterns               |
|         | **TOTAL**         | **100**  | Score >= 80 = pass                   |

---
## [4][ANTI_BLOAT]
>**Dictum:** *Prohibited patterns degrade output quality.*

<br>

**Prohibited Patterns:**
```regex
/Sourced from|Confirmed with|Based on/i  → meta-commentary
/This file|We do|You should/i            → self-reference
/might|could|probably|perhaps/i          → hedging
/\b(please|kindly|just|really)\b/i       → filler stopwords
```

[IMPORTANT] Prohibited patterns deduct 2 points per violation (maximum deduction: 10 points).

---
## [5][ERROR_SYMPTOMS]
>**Dictum:** *Symptom diagnosis accelerates fix identification.*

<br>

| [SYMPTOM]                    | [CAUSE]                      | [FIX]                                  |
| ---------------------------- | ---------------------------- | -------------------------------------- |
| Format not parsed            | Invalid JSON/YAML/XML syntax | Validate with parser before deployment |
| Score below 80               | Missing required sections    | Add all `required: true` sections      |
| Delimiter inconsistency      | Mixed fence/separator styles | Standardize throughout output          |
| Variable unresolved          | Missing default for optional | Add `${var:-default}` syntax           |
| Style duplication            | Voice rules not delegated    | Reference `style-standards` skill      |
| Scope conflict               | Precedence not respected     | Higher scope (command > global) wins   |
| Reference file not found     | Invalid `@` path             | Verify file exists in `.claude/styles/`|

---
## [6][OPERATIONAL_COMMANDS]
>**Dictum:** *Verification requires observable outcomes.*

<br>

```bash
# Format validation
jq . output.json                      # JSON syntax check
yq . output.yaml                      # YAML syntax check

# LOC verification
wc -l SKILL.md                        # Must be < 400 (full depth)
wc -l references/*.md                 # Each must be < 200

# Reference path verification
ls .claude/styles/                    # Verify format files exist
ls .claude/output-styles/             # Verify style files exist

# Anti-bloat check
grep -iE "Sourced from|Confirmed with|might|could" output.md
# Should return no matches
```

```