Skip to content

samibajwaisking/Token-Optimizer-Skill

token-optimizer

Verified

Maximize Claude's output quality while minimizing input token usage. Use this skill whenever a user wants to compress prompts, reduce token consumption, extract maximum output from Claude, write high-density instructions, optimize system prompts, or improve AI communication efficiency. Trigger on phrases like "optimize my prompt", "too many tokens", "make this shorter but better", "get more from Claude", "compress this prompt", "write a better system prompt", "token efficient", or any request to improve how someone communicates with Claude or any LLM. Also trigger when building AI-powered tools, chatbots, agents, or any system where prompt cost or quality matters.

references/output-formats.md

View on GitHub
# Output Format Templates

Explicit format instructions eliminate response bloat. These templates are ready to
append to any prompt. Choose the format that matches your use case.

---

## Format 1 — Bullet Points (General)

```
Format: bullet points
Rules: max 6 bullets, each bullet max 15 words, no sub-bullets unless critical
```

**When to use:** Quick overviews, feature lists, key takeaways, comparison highlights

---

## Format 2 — Numbered Steps

```
Format: numbered steps
Rules: action verb starts each step, one action per step, no explanatory paragraphs
```

**When to use:** Tutorials, processes, how-to guides, setup instructions

---

## Format 3 — Table

```
Format: markdown table
Columns: [col1] | [col2] | [col3]
Rules: no rows outside table, no intro paragraph, no summary after table
```

**When to use:** Comparisons, data, pricing, feature matrices, schedules

---

## Format 4 — JSON

```
Format: valid JSON only
Schema: { "key": "value" }
Rules: no markdown fences, no explanation text, parseable JSON only
```

**When to use:** API integrations, data extraction, structured output for applications

---

## Format 5 — Code Block

```
Format: code block only ([language])
Rules: no explanation unless I ask "explain:", production-ready, include comments
for non-obvious logic only
```

**When to use:** Code generation, scripts, functions, configuration files

---

## Format 6 — Structured Sections

```
Format: sections with bold headers
Sections: [Section 1] | [Section 2] | [Section 3]
Rules: each section max 3 sentences, no section longer than others by more than 50%
```

**When to use:** Reports, analyses, structured documents, multi-part answers

---

## Format 7 — One-Liner

```
Format: single sentence only
Rules: no preamble, no explanation, answer in one sentence maximum
```

**When to use:** Quick facts, definitions, yes/no questions with brief justification

---

## Format 8 — Two-Column Comparison

```
Format: two-column comparison (Option A vs Option B)
Structure:
| Aspect | Option A | Option B |
Rules: max 8 rows, include a "Winner" row at the bottom
```

**When to use:** Decision-making, product comparisons, approach tradeoffs

---

## Format 9 — Executive Summary

```
Format: executive summary
Structure:
- Bottom line: [1 sentence]
- Key points: [3 bullets max]
- Recommendation: [1 sentence]
Rules: total under 100 words, no background/context section
```

**When to use:** Business decisions, reports for non-technical audiences, briefings

---

## Format 10 — Q&A Pairs

```
Format: Q&A pairs
Structure: Q: [question] / A: [answer, max 2 sentences]
Rules: no narrative between pairs, cover [N] most important questions
```

**When to use:** FAQ generation, knowledge base content, interview prep

---

## Universal Anti-Filler Block

Add this to any prompt to eliminate all padding:

```
NO:
- Intro sentences ("Sure!", "Great question!", "Of course!")
- Restatement of my prompt
- Conclusion paragraphs ("I hope this helps!")
- Unnecessary qualifiers ("It's worth noting that...")
- Hedging unless factually required ("This may vary but generally...")
Start your response with the first word of actual content.
```

---

## Format by Output Length

| Goal | Instruction |
|---|---|
| Ultra-short | `max: 50 words` |
| Short | `max: 100 words` |
| Medium | `max: 300 words` |
| Long | `max: 600 words` |
| Comprehensive | `min: 800 words, cover fully` |
| Exact | `exactly [N] words` |

---

## Combining Format + Length + Restriction

**Most powerful formula:**
```
Format: [type] | Length: [constraint] | Rules: no filler, no rephrasing prompt,
start with [first word/section name]
```

**Example:**
```
Format: numbered steps | Length: max 8 steps | Rules: no intro, no filler,
each step starts with a verb, start with "1."
```
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall