context-optimization

Name: context-optimization
Author: muratcankoylan

79installs1.4kforks

How to install context-optimization?

muratcankoylan/agent-skills-for-context-engineering/context-optimization

$npx -y skills add muratcankoylan/Agent-Skills-for-Context-Engineering --skill context-optimization

Installs into the current project.

Summary

This skill should be used for improving context efficiency: context budgeting, observation masking, prefix or KV-cache strategy, partitioning, token-cost reduction, retrieval scoping, and extending effective context capacity without lowering answer quality.

Files · 1

View on GitHub

SKILL.md

1	# Context Optimization Techniques
2
3	Context optimization extends the effective capacity of limited context windows through strategic compression, masking, caching, and partitioning. Effective optimization increases useful capacity without requiring larger models or longer windows — but only when applied with measurement discipline. The techniques below are ordered by impact and risk.
4
5	## When to Activate
6
7	Activate this skill when:
8	- Context budgets or token costs constrain task complexity
9	- Observation masking can replace verbose tool outputs with retrievable references
10	- Prefix or KV-cache hit rate needs improvement
11	- Retrieval scoping can reduce irrelevant loaded context
12	- Context partitioning can extend effective capacity across agents
13	- Budget triggers are needed for masking, compaction, or partitioning
14
15	Do not activate this skill for adjacent work owned by other skills:
16	- Explaining why attention or context windows behave this way: `context-fundamentals`.
17	- Diagnosing active lost-in-middle, poisoning, distraction, confusion, or clash: `context-degradation`.
18	- Designing a structured handoff summary for a long conversation: `context-compression`.
19	- Storing large outputs, plans, or logs as files: `filesystem-context`.
20
21	## Core Concepts
22
23	Apply four primary strategies in this priority order:
24
25	1. KV-cache optimization — Reorder and stabilize prompt structure so the inference engine reuses cached Key/Value tensors. This is the cheapest optimization when the runtime supports prefix caching: low quality risk, immediate cost and latency savings. Apply it first when stable prefixes exist.
26
27	2. Observation masking — Replace verbose tool outputs with compact references once their purpose has been served. Tool outputs can dominate agent trajectories (claim-context-optimization-tool-output-dominance), so masking often yields the largest capacity gains. The original content remains retrievable if needed downstream.
28
29	3. Compaction — Summarize accumulated context when utilization exceeds 70%, then reinitialize with the summary. This distills the window's contents while preserving task-critical state. Compaction is lossy — apply it after masking has already removed the low-value bulk.
30
31	4. Context partitioning — Split work across sub-agents with isolated contexts when a single window cannot hold the full problem. Each sub-agent operates in a clean context focused on its subtask. Reserve this for tasks where estimated context exceeds 60% of the window limit, because coordination overhead is real.
32
33	The governing principle: context quality matters more than quantity. Every optimization preserves signal while reducing noise. Measure before optimizing, then measure the optimization's effect.
34
35	## Detailed Topics
36
37	### Compaction Strategies
38
39	Trigger compaction when context utilization exceeds 70%: summarize the current context, then reinitialize with the summary. This distills the window's contents in a high-fidelity manner, enabling continuation with minimal performance degradation. Prioritize compressing tool outputs first (they consume 80%+ of tokens), then old conversation turns, then retrieved documents. Never compress the system prompt — it anchors model behavior and its removal causes unpredictable degradation.
40
41	Preserve different elements by message type:
42
43	- Tool outputs: Extract key findings, metrics, error codes, and conclusions. Strip verbose raw output, stack traces (unless debugging is ongoing), and boilerplate headers.
44	- Conversational turns: Retain decisions, commitments, user preferences, and context shifts. Remove filler, pleasantries, and exploratory back-and-forth that led to a conclusion already captured.
45	- Retrieved documents: Keep claims, facts, and data points relevant to the active task. Remove supporting evidence and elaboration that served a one-time reasoning purpose.
46
47	Target 50-70% token reduction with less than 5% quality degradation. If compaction exceeds 70% reduction, audit the summary for critical information loss — over-aggressive compaction is the most common failure mode.
48
49	### Observation Masking
50
51	Mask observations selectively based on recency and ongoing relevance — not uniformly. Apply these rules:
52
53	- Never mask: Observations critical to the current task, observations from the most recent turn, observations used in active reasoning chains, and error outputs when debugging is in progress.
54	- Mask after 3+ turns: Verbose outputs whose key points have already been extracted into the conversation flow. Replace with a compact reference: `[Obs:{ref_id} elided. Key: {summary}. Full content retrievable.]`
55	- Always mask immediately: Repeated/duplicate outputs, b

People who liked this also viewed

6 picks

Type

context-optimization

79installs1.4kforks

View on GitHub

How to install context-optimization?

muratcankoylan/agent-skills-for-context-engineering/context-optimization

$npx -y skills add muratcankoylan/Agent-Skills-for-Context-Engineering --skill context-optimization

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Context Optimization Techniques
2
3	Context optimization extends the effective capacity of limited context windows through strategic compression, masking, caching, and partitioning. Effective optimization increases useful capacity without requiring larger models or longer windows — but only when applied with measurement discipline. The techniques below are ordered by impact and risk.
4
5	## When to Activate
6
7	Activate this skill when:
8	- Context budgets or token costs constrain task complexity
9	- Observation masking can replace verbose tool outputs with retrievable references
10	- Prefix or KV-cache hit rate needs improvement
11	- Retrieval scoping can reduce irrelevant loaded context
12	- Context partitioning can extend effective capacity across agents
13	- Budget triggers are needed for masking, compaction, or partitioning
14
15	Do not activate this skill for adjacent work owned by other skills:
16	- Explaining why attention or context windows behave this way: `context-fundamentals`.
17	- Diagnosing active lost-in-middle, poisoning, distraction, confusion, or clash: `context-degradation`.
18	- Designing a structured handoff summary for a long conversation: `context-compression`.
19	- Storing large outputs, plans, or logs as files: `filesystem-context`.
20
21	## Core Concepts
22
23	Apply four primary strategies in this priority order:
24
25	1. KV-cache optimization — Reorder and stabilize prompt structure so the inference engine reuses cached Key/Value tensors. This is the cheapest optimization when the runtime supports prefix caching: low quality risk, immediate cost and latency savings. Apply it first when stable prefixes exist.
26
27	2. Observation masking — Replace verbose tool outputs with compact references once their purpose has been served. Tool outputs can dominate agent trajectories (claim-context-optimization-tool-output-dominance), so masking often yields the largest capacity gains. The original content remains retrievable if needed downstream.
28
29	3. Compaction — Summarize accumulated context when utilization exceeds 70%, then reinitialize with the summary. This distills the window's contents while preserving task-critical state. Compaction is lossy — apply it after masking has already removed the low-value bulk.
30
31	4. Context partitioning — Split work across sub-agents with isolated contexts when a single window cannot hold the full problem. Each sub-agent operates in a clean context focused on its subtask. Reserve this for tasks where estimated context exceeds 60% of the window limit, because coordination overhead is real.
32
33	The governing principle: context quality matters more than quantity. Every optimization preserves signal while reducing noise. Measure before optimizing, then measure the optimization's effect.
34
35	## Detailed Topics
36
37	### Compaction Strategies
38
39	Trigger compaction when context utilization exceeds 70%: summarize the current context, then reinitialize with the summary. This distills the window's contents in a high-fidelity manner, enabling continuation with minimal performance degradation. Prioritize compressing tool outputs first (they consume 80%+ of tokens), then old conversation turns, then retrieved documents. Never compress the system prompt — it anchors model behavior and its removal causes unpredictable degradation.
40
41	Preserve different elements by message type:
42
43	- Tool outputs: Extract key findings, metrics, error codes, and conclusions. Strip verbose raw output, stack traces (unless debugging is ongoing), and boilerplate headers.
44	- Conversational turns: Retain decisions, commitments, user preferences, and context shifts. Remove filler, pleasantries, and exploratory back-and-forth that led to a conclusion already captured.
45	- Retrieved documents: Keep claims, facts, and data points relevant to the active task. Remove supporting evidence and elaboration that served a one-time reasoning purpose.
46
47	Target 50-70% token reduction with less than 5% quality degradation. If compaction exceeds 70% reduction, audit the summary for critical information loss — over-aggressive compaction is the most common failure mode.
48
49	### Observation Masking
50
51	Mask observations selectively based on recency and ongoing relevance — not uniformly. Apply these rules:
52
53	- Never mask: Observations critical to the current task, observations from the most recent turn, observations used in active reasoning chains, and error outputs when debugging is in progress.
54	- Mask after 3+ turns: Verbose outputs whose key points have already been extracted into the conversation flow. Replace with a compact reference: `[Obs:{ref_id} elided. Key: {summary}. Full content retrievable.]`
55	- Always mask immediately: Repeated/duplicate outputs, b

People who liked this also viewed

6 picks

Type