Shed — Context Hygiene for Agents
Shed what you don't need. Keep what matters.
Named for molting — the process of shedding an outer layer to grow. Your context window is your skin. When it gets too heavy, shed the dead weight.
Core Principle
Tool outputs are 84% of your context growth but the lowest-value tokens you carry. (Lindenbauer et al., NeurIPS 2025 DL4C workshop, measured on SWE-agent). Everything flows from this.
The Rules
After Every Tool Call
- Extract, don't accumulate. When a tool returns large output (file contents, search results, logs, API responses), immediately write the key facts to a file or compress into bullets. The raw output is now disposable.
- Ask: "Will I need this verbatim later?" Almost never. The answer you extracted is what matters, not the 500 lines that contained it.
When Context Reaches ~70%
- Trigger condensation. Don't wait for the platform to compact you — that's losing control of your own memory. At 70%, actively shed.
- Mask old tool outputs first (free, no LLM calls). Keep your reasoning and action history intact — you need your decision chain, not the raw
ls -la from 20 turns ago.
- Summarize reasoning only as backup. If masking isn't enough, compress old reasoning turns. But this is lossy and costs an LLM call — use sparingly.
- Never re-summarize a summary. If you've already condensed once and context is growing again, switch context or spawn a sub-agent. Recursive summarization compounds errors.
When Completing a Task
- Write results to file, then switch context immediately. Stale completed-task context is anti-signal for your next task. Don't carry it.
- Leave breadcrumbs. Before switching: write what you did, what's next, and where the files are to
memory/YYYY-MM-DD.md. Future-you needs a trailhead, not a transcript.
When Delegating Work
- Spawn fresh-context sub-agents for complex sub-tasks. Your context is noise for their work. Give them a clean prompt with just what they need.
- Don't inherit parent context into children. The AutoGen pattern: each agent gets its own token budget. Inherited bloat = inherited degradation.
Architecture (For Agent Builders)
- Structure context into typed blocks with hard size limits. Every production framework converges here — Letta uses labeled blocks (human, persona, knowledge) with character caps. A monolithic context is unmanageable.
- Separate working memory (in-context) from reference memory (file/DB). Your effective context is much smaller than your window size. Models lose information in the middle of long contexts.
- Place critical information at the beginning or end of context, never the middle. Positional attention bias underweights middle content by up to 15 percentage points (Hsieh et al., 2024, "Found in the Middle").
The Complexity Trap
Don't assume sophisticated compression (LLM summarization) beats simple approaches (observation masking). The JetBrains "Complexity Trap" paper (2025) tested both across 5 model configurations on SWE-bench Verified:
- Simple masking halved cost relative to raw agent
- Masking matched or exceeded LLM summarization solve rates
- Example: Qwen3-Coder went from 53.8% → 54.8% with masking alone
The lesson: start simple. Mask tool outputs. Only add summarization if masking alone isn't enough.
Cost Model
Without intervention, cost per turn scales quadratically (each turn adds tokens AND reprocesses all previous tokens). Periodic condensation converts this to linear scaling. OpenHands measured 2x cost reduction with their condenser.
Quick Reference
| Situation |
Action |
| Tool returned big output |
Extract facts → file → discard raw |
| Context at ~70% |
Mask old tool outputs |
| Context still growing after masking |
Summarize oldest reasoning turns |
| Task complete |
Write results → switch context |
| Complex sub-task needed |
Spawn fresh sub-agent |
| Already condensed, still growing |
Switch context or spawn |
| Critical info to preserve |
Put at start or end, not middle |
Sources
1---2name: shed3description: Context window hygiene for long-running LLM agents. Decision rules for when and how to compress, mask, switch, or delegate context — backed by research (JetBrains/NeurIPS 2025, OpenHands, Letta/MemGPT, LLMLingua). Use when an agent runs for extended sessions, accumulates large tool outputs, approaches context limits, or suffers from compaction/overflow. Also use when designing agent architectures that need to manage context over time.4---56# Shed — Context Hygiene for Agents78*Shed what you don't need. Keep what matters.*910Named for molting — the process of shedding an outer layer to grow. Your context window is your skin. When it gets too heavy, shed the dead weight.1112## Core Principle1314**Tool outputs are 84% of your context growth but the lowest-value tokens you carry.** (Lindenbauer et al., NeurIPS 2025 DL4C workshop, measured on SWE-agent). Everything flows from this.1516## The Rules1718### After Every Tool Call19201. **Extract, don't accumulate.** When a tool returns large output (file contents, search results, logs, API responses), immediately write the key facts to a file or compress into bullets. The raw output is now disposable.212. **Ask: "Will I need this verbatim later?"** Almost never. The answer you extracted is what matters, not the 500 lines that contained it.2223### When Context Reaches ~70%24253. **Trigger condensation.** Don't wait for the platform to compact you — that's losing control of your own memory. At 70%, actively shed.264. **Mask old tool outputs first** (free, no LLM calls). Keep your reasoning and action history intact — you need your decision chain, not the raw `ls -la` from 20 turns ago.275. **Summarize reasoning only as backup.** If masking isn't enough, compress old reasoning turns. But this is lossy and costs an LLM call — use sparingly.286. **Never re-summarize a summary.** If you've already condensed once and context is growing again, switch context or spawn a sub-agent. Recursive summarization compounds errors.2930### When Completing a Task31327. **Write results to file, then switch context immediately.** Stale completed-task context is anti-signal for your next task. Don't carry it.338. **Leave breadcrumbs.** Before switching: write what you did, what's next, and where the files are to `memory/YYYY-MM-DD.md`. Future-you needs a trailhead, not a transcript.3435### When Delegating Work36379. **Spawn fresh-context sub-agents for complex sub-tasks.** Your context is noise for their work. Give them a clean prompt with just what they need.3810. **Don't inherit parent context into children.** The AutoGen pattern: each agent gets its own token budget. Inherited bloat = inherited degradation.3940### Architecture (For Agent Builders)414211. **Structure context into typed blocks with hard size limits.** Every production framework converges here — Letta uses labeled blocks (human, persona, knowledge) with character caps. A monolithic context is unmanageable.4312. **Separate working memory (in-context) from reference memory (file/DB).** Your effective context is much smaller than your window size. Models lose information in the middle of long contexts.4413. **Place critical information at the beginning or end of context, never the middle.** Positional attention bias underweights middle content by up to 15 percentage points (Hsieh et al., 2024, "Found in the Middle").4546## The Complexity Trap4748Don't assume sophisticated compression (LLM summarization) beats simple approaches (observation masking). The JetBrains "Complexity Trap" paper (2025) tested both across 5 model configurations on SWE-bench Verified:4950- Simple masking **halved cost** relative to raw agent51- Masking **matched or exceeded** LLM summarization solve rates52- Example: Qwen3-Coder went from 53.8% → 54.8% with masking alone5354The lesson: start simple. Mask tool outputs. Only add summarization if masking alone isn't enough.5556## Cost Model5758Without intervention, cost per turn scales **quadratically** (each turn adds tokens AND reprocesses all previous tokens). Periodic condensation converts this to **linear** scaling. OpenHands measured 2x cost reduction with their condenser.5960## Quick Reference6162| Situation | Action |63|-----------|--------|64| Tool returned big output | Extract facts → file → discard raw |65| Context at ~70% | Mask old tool outputs |66| Context still growing after masking | Summarize oldest reasoning turns |67| Task complete | Write results → switch context |68| Complex sub-task needed | Spawn fresh sub-agent |69| Already condensed, still growing | Switch context or spawn |70| Critical info to preserve | Put at start or end, not middle |7172## Sources7374- Lindenbauer et al., "The Complexity Trap" (NeurIPS 2025 DL4C): https://arxiv.org/abs/2508.2143375- OpenHands Context Condensation (2025): https://openhands.dev/blog/openhands-context-condensensation-for-more-efficient-ai-agents76- Letta/MemGPT Memory Blocks: https://www.letta.com/blog/memory-blocks77- LLMLingua-2 (ACL 2024): https://aclanthology.org/2024.acl-long.91/78- Liu et al., "Lost in the Middle" (2023): https://arxiv.org/abs/2307.0317279- Hsieh et al., "Found in the Middle" (2024): https://arxiv.org/abs/2406.1600880- MEM1 Dynamic State Management (2025): https://arxiv.org/abs/2506.15841
dvcrn