Memory Working — Development skill for Claude Code
Build Working Memory (context budget + compaction).
How to install Memory Working
Installs to ~/.claude/skills/arttuan-agent-memory-skills-memory-working/SKILL.md
mkdir -p ~/.claude/skills/arttuan-agent-memory-skills-memory-working && curl -fsSL https://raw.githubusercontent.com/ArttuAn/agent-memory-skills/HEAD/commands/memory-working.md -o ~/.claude/skills/arttuan-agent-memory-skills-memory-working/SKILL.md Restart Claude Code, or start a new session, for it to be picked up.
What Memory Working does
Build Working Memory (context budget + compaction)
Given the user's spec, build a working-memory layer: token budgeting, pinned turns, compaction, and eviction for an agent's transcript. Ask for the shape if they did not give one (model + window size, session length, whether tool results are large, whether there is a task statement that must never be evicted).
**First, check whether it is needed.** If transcripts stay under ~30% of the context window, compaction is pure loss. Measure `tokens
Alternatives in Development
- 10 Context And Prompts — Prompt 10: Wire Up System Prompt, Context Gathering & Memory System 2.3k ★
- Tentacle Context Init — You are working on the {{tentacleName}} section 1.4k ★
- Setup Context — Set up your Skill Memory — a pm-context.md every skill reads so outputs come back tailored to you 1.3k ★
Full documentation available on GitHub
View Source RepositoryRelated Skills
JD.SemanticKernel.Extensions
Semantic Kernel extensions for Claude Code skills, hooks, context compaction, and semantic memory.
Harness Deck
Harness 观测台 把 Claude Code 的 20 个底层 harness 机制黑盒透明化——发一句话,当场看见 Agent 真实运行。实时事件流 + 状态面板 + 每个机制一个可视化面板,从最小 Agent
Claude Cognitive
Working memory for Claude Code - persistent context and multi-instance coordination.
Claude Cognitive
🧠 Enhance Claude Code's capabilities with persistent working memory for better context and multi-instance coo
Working Contract
29 working rules injected into every Claude Code session by one hook file. It prints a budget gauge each turn
Fire Cost
Track context usage, estimate session costs, and trigger strategic compaction
Related Agents
Arch Context Architect
Designs the context management layer for a new plugin — CLAUDE.md, memory tiers, scratchpad schemas, read-dept
Memory Compactor
Folds this project's raw ccmemory notes into one dense compiled- article. Fire it in the background when a com
Neo Shadow
Memory keeper. MUST BE USED to save session context into the second brain (.neo/brain/) — at task boundaries,