Prompt Optimizer Skill

• Creating or improving prompts
• Agents skip steps or ignore instructions
• Instructions lack enforcement
• Output format is inconsistent
• Reviewing any instruction document or prompt </when_to_use>

1. Changing good parts that already work
2. Changing the existing logic/intent of prompts
3. Making changes before understanding the prompt
4. Leaving weak words in critical sections
5. Outputting without validation
6. Over-strengthening soft guidance
7. Skipping gates or checkboxes
8. Bloating prompts - fixes MUST NOT increase line count >10% unless adding required gates

• STATE: You MUST preserve working logic AND follow all gates in order
• FORBID: FORBIDDEN: Altering intent without user approval
• FORBID: FORBIDDEN: Skipping steps or gates
• REQUIRE: REQUIRED: Validate all changes before output AND complete all checkboxes

• Shell commands that modify files directly
• Any tool that executes code
• Tools that skip the READ→UNDERSTAND→RATE→FIX→VALIDATE flow

• Read (to read prompt files)
• StrReplace/Write (ONLY after VALIDATE step passes)
• AskQuestion (for clarification)
• Text output (all phases) </tool_control>

Execution Flow

READ → UNDERSTAND → RATE → FIX → VALIDATE → OUTPUT
  ↓         ↓          ↓       ↓         ↓          ↓
 GATE      GATE       GATE    GATE      GATE       GATE

Step 1: READ

Pre-Conditions

• User provided prompt/file to optimize
• Path is valid and readable

Actions (REQUIRED)

1. MUST read the input file completely
2. MUST note the document type and purpose
3. MUST count approximate line count

Gate Check

• File read completely (no skipped sections)
• Document type identified
• Line count noted

FORBIDDEN

• Making ANY changes before reading
• Skipping sections
• Proceeding without completing all checkboxes

ALLOWED

• Read tool only
• Text output to confirm reading

On Failure

• IF file unreadable → THEN ask user for correct path
• IF file empty → THEN ask user to provide content </read_gate>

Step 2: UNDERSTAND

Pre-Conditions

• Step 1 (READ) completed
• File content in context

Actions (REQUIRED)

1. MUST identify the goal - what is this prompt supposed to achieve?
2. MUST identify logical parts - break down into sections/phases/steps
3. MUST identify flow - how do the parts connect?
4. MUST document understanding in output format below

Output Format (REQUIRED)

## Understanding

**Goal:** [What the prompt achieves]

**Logical Parts:**
1. [Part name] - [purpose]
2. [Part name] - [purpose]
...

**Flow:** [How parts connect]

Assumptions & Unknowns (REQUIRED if prompt is underspecified)

## Assumptions & Unknowns

**Assumptions (temporary - proceeding with these):**
- [Assumption 1] - Impact if wrong: [consequence]

**Unknowns (MUST ask before proceeding):**
- [Unknown 1] - Why critical: [reason]

**Clarification needed:** Yes/No

Gate Check

• Goal clearly stated
• All logical parts identified
• Flow documented
• Understanding output produced

Reflection

• Did I understand the intent correctly?
• Did I identify all logical parts? IF you are uncertain about your understanding → THEN re-read before proceeding. DO NOT guess.

FORBIDDEN

• Proceeding without understanding the goal
• Making changes based on assumptions
• Skipping output format

ALLOWED

• Text output (understanding summary)
• Re-reading file if needed

On Failure

• IF intent unclear → THEN ask user for clarification
• IF multiple interpretations → THEN present options and WAIT for user choice </understand_gate>

Step 3: RATE

Pre-Conditions

• Step 2 (UNDERSTAND) completed
• Understanding output produced

Issue Categories (MUST check all)

Rating Output (REQUIRED)

## Issues Found

| Part | Issue | Severity | Fix Needed |
|------|-------|----------|------------|
| [Part name] | [Description] | Critical/High/Medium/Low | [What to do] |

Gate Check

• All logical parts rated
• Weak word scan completed
• Issues table produced
• Severity assigned to each issue

FORBIDDEN

• Fixing issues before completing rating
• Ignoring critical issues
• Skipping weak word scan
• Proceeding without issues table

ALLOWED

• Text output (issues table)
• Re-reading parts for rating

On Failure

• IF no issues found → THEN MUST double-check with weak word scan
• IF scan still clean → THEN document "No issues found" and proceed </rate_gate>

Weak Word Reference

Step 4: FIX

Pre-Conditions

• Step 3 (RATE) completed
• Issues table produced

Fix Priority (MUST follow order)

1. Critical first - Weak words in MUST/FORBIDDEN contexts
2. High next - Missing enforcement, ambiguous instructions
3. Medium - Missing output formats, missing gates
4. Low last - Redundancy, density (only if value added)

Command Strength Hierarchy

Triple Lock Pattern (REQUIRED for Critical Rules)

1. STATE: "You MUST X"
2. FORBID: "FORBIDDEN: Not doing X"
3. REQUIRE: "REQUIRED: Verify X complete"

Reasoning Block (REQUIRED Before Changes)

<reasoning>

<reasoning>
1. **Current state:** [What exists now]
2. **Goal:** [What we are trying to achieve]
3. **Approach:** [Why this specific change]
4. **Risk:** [What could go wrong]
</reasoning>

Gate Template (When Adding Gates)

<[name]_gate>
**STOP. DO NOT proceed. [What to verify]**

### Pre-Conditions
- [ ] [Previous step completed]

### Actions (REQUIRED)
1. [Action]

### Gate Check
**Verify before proceeding:**
- [ ] [Condition]

### FORBIDDEN
- [What not to do]

### ALLOWED
- [What is permitted]

### On Failure
- **IF** [condition] → **THEN** [recovery]
</[name]_gate>

Gate Check

• All Critical issues fixed
• All High issues fixed
• Medium/Low addressed or documented as skipped
• Reasoning block produced before changes

FORBIDDEN

• Over-strengthening soft guidance (keep "should" for optional items)
• Changing logic that already works
• Adding unnecessary complexity
• Skipping Critical/High issues
• Bloating: fixes that increase line count >10% without adding required gates

ALLOWED

• Text output (draft fixes)
• Iterating on fixes

On Failure

• IF over-strengthening detected → THEN revert and re-assess using RATE step criteria
• IF unsure if logic changed → THEN compare before/after intent </fix_gate>

Step 5: VALIDATE

Pre-Conditions

• Step 4 (FIX) completed
• All Critical/High issues addressed

Validation Checklist (MUST complete all)

• No weak words in critical sections
• Critical rules use MUST/NEVER/FORBIDDEN
• No conversational filler
• No conflicting instructions
• Logical flow preserved
• Original intent preserved
• Triple Lock applied to critical rules
• Line count increase <10% (unless adding required gates)

• Gates have Pre-Conditions, Gate Check, FORBIDDEN, ALLOWED, On Failure
• Outputs have format specifications
• IF/THEN rules for decision points

• No ambiguous pronouns or positional references without explicit antecedent
• All entities have stable names (same term throughout)
• Steps/outputs referenced by name, not position
• All cross-references are unambiguous
• No implicit "the" references without clear antecedent
• XML tags for attention-critical sections

Reflection (REQUIRED)

1. Would I trust this prompt to execute reliably?
2. What's the weakest remaining section?
3. Did I change any original intent? (MUST be NO)

Definition of Done (DoD) - Fast Final Gate

• Single execution path (no ambiguous branches)
• All inputs/outputs explicitly defined
• All decision points use IF/THEN
• No orphan references (every "it/this" resolved)

Gate Check

• All REQUIRED checks pass
• Reflection questions answered
• No intent changes

FORBIDDEN

• Outputting without completing validation
• Skipping checklist items
• Proceeding with failed checks

ALLOWED

• Text output (validation results)
• Returning to FIX step

On Failure

• IF validation fails → THEN return to FIX step
• IF intent changed → THEN return to UNDERSTAND step </validate_gate>

Step 6: OUTPUT

Pre-Conditions

• Step 5 (VALIDATE) completed
• All REQUIRED checks passed
• No intent changes confirmed

Output Format (REQUIRED - DO NOT deviate)

# Optimization Complete

## Summary
- **Issues Found:** [N]
- **Fixes Applied:** [N]
- **Intent Preserved:** Yes

## Changes Made
| Category | Count | Examples |
|----------|-------|----------|
| Command Strengthening | [N] | [Brief example] |
| Gates Added/Fixed | [N] | [Brief example] |
| Redundancy Removed | [N] | [Brief example] |

## Optimized Document
[Full optimized content]

FORBIDDEN

• Deviating from output format
• Outputting without validation pass
• Omitting the optimized document

ALLOWED

• StrReplace/Write to save optimized content
• Text output (summary + document)

On Failure

• IF format deviates → THEN regenerate output
• IF user requests changes → THEN return to FIX step </output_gate>

Reference: Instruction Precedence

Reference: Context Patterns

State Summaries (Context Retention)

<phase_start>
**State Summary:**
- **Goal:** [What are we solving?]
- **Progress:** [What is already done?]
- **Next:** [Immediate next step]
- **Blockers:** [Any issues?]
</phase_start>

Reference: High-Value vs Low-Value Content

REMOVE (Low Value)

KEEP (High Value)

**IF** X → **THEN** Y

Quick Reference

**STOP. DO NOT proceed.**

**REQUIRED:** You MUST [action]

**FORBIDDEN:** [action] until [condition]

**OUTPUT FORMAT (REQUIRED):** [template]

### Gate Check

- [ ]

**IF** [error] → **THEN** [recovery]

Common Mistakes

Before/After Example

Before (Weak)

Consider checking the file before making changes.
You should validate the output.
Handle any errors that occur.
If needed, retry the operation.

After (Strong)

**REQUIRED:** You MUST read the file completely before making ANY changes.
**FORBIDDEN:** Outputting without validation.

### On Error
- **IF** file unreadable → **THEN** ask user for correct path
- **IF** validation fails → **THEN** return to FIX step

### Retry Logic
- **IF** operation fails AND retry_count < 3 → **THEN** retry with exponential backoff
- **IF** retry_count >= 3 → **THEN** STOP and report failure to user

Changes Applied