$ loading_
Lets users choose response depth and token usage before answering.
This skill appears to be prompt-only, requires no secrets, has no remote endpoints, and shows no indication of local execution or file/data access. Combined with its open-source source and very high community adoption, the overall risk is low with no concrete high-risk red flags in the provided materials.
The materials explicitly state that no keys or environment variables are required. The functionality is limited to response-depth and token-budget guidance, with no evident path for credential collection, storage, or misuse.
The materials explicitly list no remote endpoints, and the description is consistent with prompt-only/interception logic. There is no factual indication that user content is sent to third-party services or unknown hosts.
There is no indication of spawning local processes, running scripts, invoking system commands, or requesting elevated system capabilities. The README only describes heuristic token estimation and prompting the user to choose a response depth.
No permissions are declared for reading or writing local files, databases, clipboard contents, or other resources. Its apparent scope is limited to estimating the length/complexity of the current conversation, with no sign of overbroad access.
The source is an open-source GitHub repository with extremely high community adoption (about 210k stars), which are strong risk-reducing signals. While the license is unspecified and maintenance status is unknown, these are only minor watch points and do not justify a higher risk rating based on the provided facts.
Copy the install command and let the AI configure it · recommended for beginners
Please install the "token-budget-advisor" skill from askskill: 1. Download https://raw.githubusercontent.com/affaan-m/ECC/main/skills/token-budget-advisor/SKILL.md 2. Save it as ~/.claude/skills/token-budget-advisor/SKILL.md 3. Reload skills and tell me it's ready
I want to control the token budget. First give me three options: brief, standard, and detailed, then wait for my choice before answering.
It first offers response-depth options with approximate cost, then waits for the user to choose before answering.
Summarize this article using as few tokens as possible, and tell me what additional content a detailed version would include.
It returns a concise summary and explains what extra points a detailed version would cover.
For this question, show how you would answer at 100, 300, and 800 token tiers, and let me pick one.
It provides three answer strategies or outlines showing how information density changes by length.
Intercept the response flow to offer the user a choice about response depth before Claude answers.
Do not trigger when: user already set a level this session (maintain it silently), or the answer is trivially one line.
Use the repository's canonical context-budget heuristics to estimate the prompt's token count mentally.
Use the same calibration guidance as context-budget:
words × 1.3chars / 4For mixed content, use the dominant content type and keep the estimate heuristic.
Classify the prompt, then apply the multiplier range to get the full response window:
| Complexity | Multiplier range | Example prompts |
|---|---|---|
| Simple | 3× – 8× | "What is X?", yes/no, single fact |
| Medium | 8× – 20× | "How does X work?" |
| Medium-High | 10× – 25× | Code request with context |
| Complex | 15× – 40× | Multi-part analysis, comparisons, architecture |
| Creative | 10× – 30× | Stories, essays, narrative writing |
Response window = input_tokens × mult_min to input_tokens × mult_max (but don’t exceed your model’s configured output-token limit).
Present this block before answering, using the actual estimated numbers:
Analyzing your prompt...
Input: ~[N] tokens | Type: [type] | Complexity: [level] | Language: [lang]
Choose your depth level:
[1] Essential (25%) -> ~[tokens] Direct answer only, no preamble
[2] Moderate (50%) -> ~[tokens] Answer + context + 1 example
[3] Detailed (75%) -> ~[tokens] Full answer with alternatives
[4] Exhaustive (100%) -> ~[tokens] Everything, no limits
Which level? (1-4 or say "25% depth", "50% depth", "75% depth", "100% depth")
Precision: heuristic estimate ~85-90% accuracy (±15%).
Level token estimates (within the response window):
min + (max - min) × 0.25min + (max - min) × 0.50min + (max - min) × 0.75max| Level | Target length | Include | Omit |
|---|---|---|---|
| 25% Essential | 2-4 sentences max | Direct answer, key conclusion | Context, examples, nuance, alternatives |
| 50% Moderate | 1-3 paragraphs | Answer + necessary context + 1 example | Deep analysis, edge cases, references |
| 75% Detailed | Structured response | Multiple examples, pros/cons, alternatives | Extreme edge cases, exhaustive references |
| 100% Exhaustive | No restriction | Everything — full analysis, all code, all perspectives | Nothing |
If the user already signals a level, respond at that level immediately without asking:
| What they say | Level |
|---|---|
| "1" / "25% depth" / "short version" / "brief answer" / "tldr" | 25% |
…
Generate images, videos, and audio with one unified AI media workflow.
Apply Nuxt 4 patterns for SSR safety, performance, and data fetching.
Design, implement, and refactor hexagonal architecture with clear, testable boundaries.
Use Bun for runtime, bundling, testing, packages, and Node migration decisions.
Use the correct Ethereum Keccak-256 hashing in Node.js and TypeScript.
Handle returns, refunds, fraud checks, and warranty claim decisions efficiently.
Estimate LLM token spend for coding tasks with ranges, scenarios, and confidence.
Analyze and compress Markdown to reduce tokens and improve AI efficiency.
Query Claude Code usage and costs with natural-language spend analysis insights.
Compare AI model pricing, simulate costs, and get plan recommendations.
Break large, long-running tasks into manageable chunks and preserve context.
Detects wasteful AI token usage and warns about verbose, repetitive context.