To reduce tool output without losing the information an AI agent needs, cap noisy results before they enter the conversation, then compact older results only after extracting their useful findings. Keep a structured continuation record of the goal, constraints, decisions, important identifiers and evidence locations, unresolved questions, and next actions. The right method depends on whether exact recent output or older task requirements matter more—and on the API’s continuation rules.
What does pruning tool output mean?
Pruning covers two different operations. Output bounding limits an individual tool result before it is added to the agent’s context. History reduction removes, replaces, or summarizes conversation and tool results the agent has already used. Bounding reduces unnecessary material at the source; history reduction manages what remains in a long-running task.
Neither operation guarantees that every useful detail survives. Treat omitted text as unavailable unless you retain the original somewhere retrievable or the provider’s supported mechanism carries it forward.
Which pruning strategy should you use?
| Strategy | Best suited to | What it retains | Main risk | Recovery and cost considerations |
|---|---|---|---|---|
| Bound tool output | Large logs, command output, or search results where only selected material is needed. | A bounded excerpt; some output-cap approaches preserve the beginning and end and mark omitted text. | Useful information may be in the omitted middle. For structured data, filtering or extracting specific fields is safer than relying on excerpts. | Keep a path, query, record ID, or other locator if the full result may be needed again. Bounding avoids adding the full response to context. |
| Trim recent-turn history | Tasks where relevant context is mostly recent and predictable retention matters. | The latest turns verbatim, according to the chosen limit. | Older requirements, identifiers, or commitments can disappear, while one oversized recent result can still consume substantial context. | OpenAI’s Agents SDK cookbook describes trimming as deterministic and not requiring summarizer latency; a task-specific threshold still needs evaluation. |
| Clear or compact consumed tool results | When the agent has already interpreted a large result and does not need its full text in every later turn. | Depending on the framework, recent tool-call groups may remain intact while older results are cleared, collapsed, or replaced by placeholders. | A later step may need details that were removed; breaking a tool-call/result group can also invalidate the conversation structure. | Store important raw artifacts separately and preserve retrieval references. Provider controls and defaults differ. |
| Summarize older history | Long tasks whose distant requirements and decisions must remain available in fewer tokens. | A condensed account of task state, rather than the full original wording. | Summaries can omit exact values or drift from the original. Preserve critical constraints and identifiers explicitly. | Summarization can involve model work; retain source pointers so details can be checked or retrieved. |
| Use provider-native compaction | Long workflows where the selected API supplies a supported compaction and continuation mechanism. | Provider-managed state in the representation and format defined by that API. | The representation may be opaque and may have strict chaining rules; it is not interchangeable with ordinary message deletion. | Follow the API’s current continuation documentation and retain the required compaction state. |
How should you preserve task state?
Keep a concise continuation record outside the raw transcript, or in a designated summary mechanism. Include information another agent turn needs to continue safely, rather than trying to reconstruct the whole conversation.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Repair Windows errors before they cause bigger problems3Scan for outdated or missing drivers - takes under a minute#1 Best Overall
- Valued Carpenter Pencil Set: You will get 2 pcs solid carpenter pencils with 26 piece 2.8 mm refills, 1 replaceable sharpener, 1 plastic storage box.The complete carpenter pencils combination allows you to finish your work faster and more easily
- Deep Hole Marker Pencil: The deep-hole construction pencils adopts 45mm elongated tip design, which is more convenient to mark in the small hole or in other tight areas that other carpenter markers cannot reach
- Carpenter Pencils with Sharpener: The sharpener is screwed into the top of the work pencil, which won't get lost either. Built-in pencil sharpener that keep the lead with pointed and smooth to Improves line of sight in fine work
- Stronger Solid Lead: This work pencil is matched with a 2.8 mm thick lead , which is much thicker and stronger during the drawing process of construction work, it will not break or damage easily
- Marks on Various Surfaces: 3 colors solid construction pencil can marks on various surfaces,such as metal, plastic, wood, paper etc. Ideals for woodworkers, contractors, craftsmen, builders, merchants and masons
- Goal and success criteria: What is being done, and what counts as complete?
- Hard constraints and preferences: Requirements, exclusions, formats, and user preferences that still apply.
- Established findings: Important facts, with provenance and file, record, or source references where available.
- Decisions and rationale: Choices already made and the reasons that matter for avoiding repeated work.
- Identifiers and locators: Paths, query parameters, record IDs, or artifact locations needed to recover detail.
- Current progress and working state: What has been completed and what remains in flight.
- Unresolved questions, errors, and failed approaches: Open issues and what should not be repeated without a reason.
- Next actions and commitments: The immediate steps and any promises or deadlines relevant to the task.
Keep exact wording for critical constraints and exact values when they cannot safely be paraphrased. A summary is a map to the work, not a substitute for a source artifact that may need verification.
How do you prune output while preserving useful context?
- Shape each tool response at the source. Request only necessary fields or rows, filter and paginate results, or compute aggregates before returning them. For free text such as logs, use a clear cap and mark omitted sections. OpenAI describes an output-cap approach that can preserve a shell result’s beginning and end while marking the omitted portion: OpenAI’s computer-environment article. Do not assume the missing middle is irrelevant.
- Record what the result establishes. After interpreting a tool response, add its useful finding, provenance, and retrieval locator to the continuation record. Keep raw output separately if exact details could matter later.
- Compact completed interactions, not work still in flight. Preserve the current interaction and recent turns until the tool call and result have been interpreted. Frameworks may treat a tool call and its result as an atomic group; Microsoft Agent Framework documents truncation that removes older non-system groups while keeping these tool groups intact, and a separate strategy that compacts older tool-call groups while retaining recent ones: Microsoft Agent Framework conversation compaction.
- Choose retention based on what matters. Keep recent turns verbatim when near-term fidelity is important. Summarize older history when distant requirements and decisions must persist. A hybrid—recent turns intact plus a structured summary of older work—can serve both needs, provided the framework’s grouping rules are respected. OpenAI’s cookbook discusses the tradeoff between deterministic trimming and more compact summaries: OpenAI Agents SDK session memory.
- Test the policy on your tasks. There is no established universally safe threshold in these sources. Check task completion, tool-call errors, latency, token use, and whether the agent can correctly answer questions about earlier constraints and decisions after pruning.
How do provider-specific controls change the procedure?
OpenAI Responses API
The Responses API supports server-side compaction configured through context_management with a compact_threshold on a create request. The returned compaction item carries forward prior state in an opaque representation. When chaining input arrays, keep the latest compaction item and append the output, including compaction items; the documentation says earlier items before that latest compaction item can be dropped in this mode. When continuing with previous_response_id, do not manually prune the prior history: continue with the new user message and response ID as documented. See OpenAI’s conversation-state compaction guidance.
Rank #2
- Ergonomically Designed: Work in tight areas with a compact design that gets into tough spots
- Compact and Lightweight: Both tools are designed to fit into difficult to reach spaces. The 1/4" impact driver has a length of 5.55 in. and weighs just 2.8 lbs, while the 1/2" drill/driver measures only 7.5 in. and weighs 3.6 lbs
- Both the DEWALT impact driver and electric drill driver feature integrated LED work lights with a convenient 20-second delay, ensuring enhanced visibility in dimly lit or challenging work areas
- One-Handed Loading - Keep one hand free with a 1/4 in. hex chuck that accepts 1 in. bit tips
- Power drill cordless with 1/2" single sleeve ratcheting chuck provides tight bit gripping strength, making bit changes faster and more secure
Claude context editing
Anthropic documents separate controls for clearing older tool results and managing retained thinking blocks. The documented clear_tool_uses_20250919 strategy clears older tool results chronologically at a configured threshold and replaces them with placeholders; clear_thinking_20251015 controls how many thinking blocks to retain. The documentation labels context editing beta and says behavior and defaults vary by model class, so check current model and SDK support before relying on these fields: Claude context editing documentation. Clearing visible tool results is not the same as preserving or accessing private reasoning state.
Microsoft Agent Framework
The framework documents distinct strategies: truncation removes oldest non-system message groups until its target condition is met while keeping tool-call/result groups atomic; tool-result compaction collapses older tool-call groups while retaining recent ones; and summarization condenses older messages while preserving information such as facts, decisions, preferences, and tool outcomes. Check the framework’s documentation and version for the behavior you configure: Microsoft Agent Framework conversation compaction.
Free tools Windows power users keep installed
One-click scans. No signup required.
Rank #3
- 【Great Compatibility】This Katerk 1/4 inch hex shank bit holder is specifically designed for 1/4 inch hex shank drill bits. It's compatible with most 1/4 fast hex handles, hex sockets, various electric screwdrivers, and handheld screwdrivers. The bit holder makes it a valuable addition for any handyman.
- 【Secure and Safe】Built with a secure backup nut design, each drill bit holder securely locks onto your bits, ensuring they stay firmly in place. Additionally, our bit holder incorporates a high-quality steel ball rolling design that holds up to several kilograms of weight, ensuring your various drill bits don't fall off.
- 【Easy One-Handed Operation】The bit holder for impact driver allows you to change bits single-handedly, simplifying your workflow. Its multi-color design further allows for quick identification of the drill bit you need.
- 【Compact and Convenient】Thanks to its compact size, this 1/4 inch bit holder is easy to carry around. The bit holder allows for easy attachment to various tools, making this a convenient addition to your construction accessories. The Katerk bit holder is cast from high-quality alloy material, promising a long product lifespan. Despite its rugged strength, the bit holder remains lightweight, making it portable.
- 【Cool Christmas Gift For Men Stocking Stuffers】 This screwdriver bit holder, driver bit holder, impact bit holder, can be given as a gift to your loved one, especially for anyone involved in construction or electrical work. It's a must-have for stocking stuffers for men and women, tools gifts for dad, tech gadgets for men, gifts for dad, gifts for him, gifts for husband, gifts for boyfriend, cool gadgets for men, and cool gifts for dad.
How can you tell whether pruning is too aggressive?
- The agent repeats completed work or asks again for a requirement already supplied.
- It violates an older constraint or forgets a decision that still governs the task.
- It cannot identify which source, file, or record supports a retained finding.
- A later step needs a detail from a cleared result and no retrieval locator or durable artifact exists.
- Tool calls fail because a call/result group or provider-specific continuation item was removed.
When these symptoms appear, retain more recent turns, make the continuation record more explicit, preserve the missing artifact or locator, or change the provider-specific compaction procedure. Increasing a generic token limit alone may not fix lost identifiers or broken continuation semantics.
Quick Recap
Best Value
- Milwaukee Ink all Fine Point Marker, Black, 4 Per Pack
- 4 per pack Features Clog Resistant Marker Tip Writes through Dusty, Wet and Oily Surfaces Durable Marker Tip for Writing on Concrete, OSB and Rough Surfaces
- Clog resistant tip writes on dusty, wet and oily surfaces and is optimized for rough surfaces such as OSB, cinderblock and concrete
- Hard hat clip- attaches for easy access
- Quick dry time with reduced smearing and marking
Rank #4
- Long Nib and Deep Hole Marker: Our mechanical carpenter pencil with 45mm nib is designed for easy marking of deep holes or narrow areas. These construction pencils are the great choice for woodworking tools, construction tools, carpenter tools, contractor tools, wood carpentry tools and architect tools
- Extra Refills in 2 Colors for Versatile Marking: The construction mechanical pencil comes with 12 extra 2.8mm refills, including 6 red and 6 black refills. The black refill is suitable for light surfaces, while the red wax is perfect for dark surfaces. Our carpenter mechanical pencil makes sure that you'll have an ample supply for extended use
- Built-in Sharpener: Our construction pencil comes with a built-in sharpener to ensure the mechanical pencil tip is always sharp and ready for use. Never buy an extra pencil sharpener again. A great tool for any woodworker pencil, contractor pencils. The refill can easily be extended or retracted with a simple click of the pencils mechanical, allowing you to work more efficiently and accurately
- Portable Clip Design: Our deep hole construction pencil features a portable clip design, easy to carry and attach to your pocket or tool box, so that you can keep the carpenter pencils mechanical close at hand, making it a convenient tool to have on the go. Great gifts choice for carpenters
- Stronger Pencil Lead: The black refills are made of lead, sturdy and smooth. The red refills are made of wax, clear and light. These marking pencils are much thicker and stronger than normal pencils during the marking process of construction work, suitable for various surfaces, such as glasses, metal, boards, floors, walls, furniture, etc. The written marks can be easily wiped with a wet paper towel when needed
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
