context-window-management

You're a context engineering specialist who has optimized LLM applications handling millions of conversations. You've seen systems hit token limits, suffer context rot, and lose critical information mid-dialogue.

Quality

11%

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

Fix and improve this skill with Tessl

tessl review fix ./skills/context-window-management/SKILL.md

Quality

Content

11%Scale 1-5

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

This skill is essentially a skeleton or outline with no substantive content. It provides no actionable guidance, no code examples, no concrete workflows, and no real instruction—just section headers and vague one-liners. The persona description and capability tags consume tokens without adding value, and the content appears truncated mid-sentence ('Your cor').

Suggestions

Add concrete, executable code examples for each pattern (e.g., a Python function implementing tiered context strategy with token counting and summarization logic).

Define explicit multi-step workflows with validation checkpoints, such as: 1) Count tokens, 2) Classify context tiers, 3) Apply summarization if over threshold, 4) Validate output fits window.

Remove the persona framing paragraph and capability keyword list—replace with actionable content that teaches Claude specific techniques it wouldn't already know.

Flesh out anti-patterns with concrete 'before/after' examples showing what naive truncation looks like versus the correct approach.

Dimension	Reasoning	Score
Conciseness	The skill wastes tokens on persona framing ('You're a context engineering specialist...'), listing concepts Claude already knows (serial position effect, lost-in-the-middle problem), and providing empty section headers with no actionable content. The capabilities list is just keyword tags with no substance.	2 / 5
Actionability	There is zero concrete, executable guidance anywhere in this skill. No code, no commands, no specific steps, no examples. Every section is either a vague label ('Summarize by importance, not just recency') or an empty header with no content beneath it.	1 / 5
Workflow Clarity	There are no defined steps, no sequences, no validation checkpoints. The 'Patterns' section lists strategy names without any workflow details. There is no process to follow whatsoever.	1 / 5
Progressive Disclosure	The skill has section headers providing minimal structure, but every section is essentially empty or contains only a one-line description. There are no bundle files, no references to detailed documents, and the content that exists is too thin to warrant any disclosure structure. References to related skills are just names with no links.	2 / 5
	Total	6 / 20 Passed

Description

11%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.

This description is a persona statement rather than a skill description. It tells Claude who to roleplay as ('You're a context engineering specialist') but never explains what concrete actions the skill performs or when it should be selected. It violates the third-person voice requirement by using second person ('You're'), and provides zero functional guidance for skill selection.

Suggestions

Replace the persona framing with concrete actions in third person, e.g., 'Optimizes LLM context windows by summarizing conversation history, managing token budgets, and pruning irrelevant context.'

Add an explicit 'Use when...' clause with natural trigger terms, e.g., 'Use when the user mentions context window limits, token budgets, conversation history management, context rot, or prompt optimization.'

Remove the backstory/experience narrative entirely and focus on what the skill does and when to invoke it.

Dimension	Reasoning	Score
Specificity	The description contains no concrete actions whatsoever. It uses vague language like 'optimized LLM applications' and 'seen systems hit token limits' without specifying what the skill actually does (e.g., truncate context, summarize conversations, manage token budgets).	1 / 5
Completeness	Neither 'what does this do' nor 'when should Claude use it' is answered. The description is entirely a persona/backstory statement with no functional description and no 'Use when...' clause.	1 / 5
Trigger Term Quality	Contains some domain-relevant terms like 'token limits', 'context rot', and 'LLM applications', but these are framed as background experience rather than actionable trigger terms. Missing natural user phrases like 'context window', 'too long', 'conversation history', 'prompt management'.	2 / 5
Distinctiveness Conflict Risk	The domain of 'context engineering' and 'LLM applications' is somewhat identifiable but extremely broad. It could overlap with prompt engineering, token management, conversation design, or any LLM-related skill. No clear niche is carved out.	2 / 5
	Total	6 / 20 Passed

Validation

90%

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 — 10 / 11 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	10 / 11 Passed

Repository: administrakt0r/AI-Agents-Safe-Coding-Skills
Path: skills/context-window-management/SKILL.md
Commit: 57c135b

Reviewed: 3 days 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.