nowait-reasoning-optimizer

Implements the NOWAIT technique for efficient reasoning in R1-style LLMs. Use when optimizing inference of reasoning models (QwQ, DeepSeek-R1, Phi4-Reasoning, Qwen3, Kimi-VL, QvQ), reducing chain-of-thought token usage by 27-51% while preserving accuracy. Triggers on "optimize reasoning", "reduce thinking tokens", "efficient inference", "suppress reflection tokens", or when working with verbose CoT outputs.

Quality

83%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

71%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is well-structured with executable integration examples and genuinely useful technique-specific guidance, but it suffers from repeated RL-vs-distilled messaging and a broken reference to a missing `references/keywords.md` file.

Suggestions

Create the missing `references/keywords.md` file (or remove the reference and inline the complete keyword list) so the signaled navigation target actually exists.

Consolidate the RL-vs-distilled guidance into a single section to remove the triple repetition across 'Supported Models', 'RL vs Distilled Models', and the explanation paragraph.

Add a brief verification step (e.g., comparing token counts or spot-checking accuracy on a sample) so users can confirm NOWAIT is behaving as expected after integration.

Dimension	Reasoning	Score
Conciseness	Mostly efficient with useful technique-specific knowledge, but the RL-vs-distilled caveat is repeated three times (Supported Models note, RL vs Distilled table, and explanatory paragraph) and two separate reduction tables overlap, so it could be tightened.	3 / 5
Actionability	Provides fully executable, copy-paste-ready code for the common cases: a HuggingFace Transformers integration and a vLLM integration, both importing the bundled processor and showing complete generation calls.	5 / 5
Workflow Clarity	The single apply-the-processor action is unambiguous and shown in code, and model-selection guidance (RL vs distilled) is clear; minor gap is the absence of any verification step for confirming output quality after applying NOWAIT.	4 / 5
Progressive Disclosure	Sections are well-organized and references are clearly signaled at one level deep, but `references/keywords.md` is referenced as the complete keyword list while the `references/` directory does not exist, so a signaled reference leads nowhere.	3 / 5
	Total	15 / 20 Passed

Description

95%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is third-person, concise, and clearly states both what the skill does and when to use it, with concrete trigger phrases and quantified outcomes. It is a strong, well-targeted description with only minor room for broader action coverage.

Dimension	Reasoning	Score
Specificity	Lists several concrete actions ('Implements the NOWAIT technique', 'reducing chain-of-thought token usage by 27-51%', 'preserving accuracy', 'suppress reflection tokens') with quantified results, though the actions cluster around a single technique so coverage is strong but not maximally comprehensive.	4 / 5
Completeness	Explicitly answers both 'what' (implements NOWAIT, reduces CoT tokens 27-51% while preserving accuracy) and 'when' ('Use when optimizing inference of reasoning models...' plus an explicit 'Triggers on' clause with concrete phrases).	5 / 5
Trigger Term Quality	Provides comprehensive natural trigger phrases users would actually say ('optimize reasoning', 'reduce thinking tokens', 'efficient inference', 'suppress reflection tokens', 'verbose CoT outputs') plus concrete model names.	5 / 5
Distinctiveness Conflict Risk	Occupies a clear niche (NOWAIT intervention for R1-style reasoning LLMs) with named model families and distinctive trigger phrases, making conflict with unrelated skills minimal.	5 / 5
	Total	19 / 20 Passed

Validation

93%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
referenced_paths_exist	Referenced path issues: 2 missing	Warning

	Total	15 / 16 Passed

Repository: foryourhealth111-pixel/Vibe-Skills
Path: bundled/skills/nowait-reasoning-optimizer/SKILL.md
Commit: d5ae560

Reviewed: 1 day ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.