
Context Preservation
Saves key values from tool results to notes before the agent's context gets pruned
- Repository
- aden-hive/hive
- Stars
- 11,132
- Licence
- Apache-2.0
- Files
- 1 file
What it does
This skill teaches an AI agent that its working memory is limited, and older tool outputs get deleted automatically. Once installed, the agent copies the important facts it will need later, such as URLs, specific values, IDs, and conclusions, into short working notes as it goes, instead of relying on long raw outputs that may be removed. It also tells the agent to pass state between tasks through the project's progress database rather than loose text blobs.
Great for
- Developers building long-running multi-step agents in the Hive framework
- Anyone whose agent loses track of earlier results during long research or data-gathering tasks
Skip it if
- Short, single-step conversations where context never fills up, since the extra note-taking adds no value
- Users outside the Hive framework, because it references Hive-specific names like progress.db and _working_notes that other setups will not have
Try asking
- “Run a multi-step research task and keep the key sources and figures noted as each search finishes”
- “Have an agent process a long batch of API results and save only the relevant fields”
- “Hand off a partly finished project between worker agents using the task database”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 1 original files, the Apache-2.0 licence and an attribution note.
context-preservation.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip context-preservation.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip context-preservation.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Expect the skill to be most useful in long tasks; it is designed to work with Hive's context-warning alerts
- Make sure the Hive task database schema is set up, because handoffs depend on the steps, tasks and sop_checklist tables it references
What's inside
- SKILL.md
Redistributed unmodified from aden-hive/hive under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.