context-compression-claude-code

Use this skill whenever the conversation context is getting long, when a user asks to "compress", "summarize", or "clean up" the conversation, or when you de...

Security Scan

The skill's instructions mostly match a conversation-compression purpose, but it explicitly tells the agent to persist 'key credentials or config' and to write structured memory files under the user's home — behavior that is disproportionate to simple compression and could lead to sensitive data being stored without clear safeguards.

Purpose & Capability

Name/description promise: compress conversation context and produce a memory file. Actual requirements: none (instruction-only). However, the skill explicitly recommends preserving 'Key credentials or config' in Layer 1 and writing those to a persistent memory file under ~/.claude/..., which is not obviously necessary for a compression utility and expands the skill's scope into storing potentially sensitive secrets.

Instruction Scope

SKILL.md instructs the agent to scan the entire conversation, generate compressed summaries, and update a persistent memory file (memory-template.md) under the user's home directory. It also provides optional hook setup that edits ~/.claude/settings.json to run a command on PreCompact. Writing conversation-derived data (and especially credentials/config) to disk and modifying hook settings is beyond a minimal 'summarize' action and increases privacy risk.

Install Mechanism

Instruction-only skill with no install spec, no downloaded code, and no extra binaries required — low installation risk. The provided setup-hook uses existing system files and a local python invocation, but nothing is downloaded or executed from remote sources.

Credentials

The skill requires no environment variables or external credentials, which is good, but it instructs preserving and storing 'Key credentials or config (accounts, paths, special settings)' in a persistent memory file. Requesting or recommending persistent storage of secrets is disproportionate to context compression and could lead to unintentional credential retention/exfiltration if the agent or environment is compromised.

Persistence & Privilege

The skill isn't 'always: true' and is user-invocable by default. However, the optional setup-hook instructions modify user hook configuration to auto-trigger compression, which would give the skill persistent automatic invocation if the user applies those changes. That behavior is optional but significant if enabled.

What to consider before installing

This skill mostly does what its name claims (compress conversation context), but it recommends persisting 'key credentials or config' into a memory file and instructs editing user hook configuration to auto-trigger — both increase risk. Before installing or enabling hooks: 1) Do not store secrets (passwords, API keys, tokens, private paths) in the memory file — redact them or omit Layer 1 credential entries. 2) Inspect and back up any config files (e.g., ~/.claude/settings.json) before applying hook changes. 3) If you allow the skill to write memory, ensure the memory file path and file permissions are restricted to your user. 4) Prefer manual invocation until you trust the exact behavior; ask the author to clarify why credentials must be preserved and what protections are used. 5) If you need help, test in an isolated account or environment first. These steps will reduce the chance sensitive data is retained or automatically processed.

Origin: This skill was extracted from Claude Code's internal implementation and rules. Claude Code openly exposes its safety mechanisms (hooks, system prompts, skill definitions) in the ~/.claude/ directory. The tiered compression strategy, layer classification system, and memory file structure were reverse-engineered from Claude Code's PreCompact/PostCompact hooks and session memory handling, then adapted for OpenClaw's multi-channel environment.

Context Compression

Why Tiered Compression Is Needed

A one-size-fits-all approach to dropping context either leaves the agent without memory or burns through context too quickly. The core principle of tiered strategy is: different information has different lifecycles. A user's statement "I prefer concise responses" is worth remembering forever; but a resolved error message from three days ago has no value today.

---

Before Compression: Identify the Scenario

Before starting compression, determine which conversation scenario you're in, as retention strategies differ:

Scenario	Characteristics	Compression Tendency
Task-oriented	Clear goal, step-driven	Keep goal and incomplete steps, compress process details
Chat-oriented	Open topics, no clear task	Keep only user preference signals, discard aggressively
Research-oriented	Gathering information, continuous accumulation	Keep conclusions and sources, compress procedural discussions
Group-chat	Multiple people, high noise	Discard aggressively, keep only directly relevant content

---

Three-Layer Compression Strategy

Layer 1: Must Preserve (Keep Original)

These contents remain in original or near-original form after compression:

• User's explicitly stated goals, requirements, deadlines
• Incomplete tasks (in-progress, interrupted)
• Confirmed important decisions ("We decided to go with Plan B")
• User-expressed explicit preferences ("I don't like X", "Always use Y format going forward")
• Key credentials or config (accounts, paths, special settings)

Layer 2: Compress to Summary (Extract and Keep)

Keep conclusions, discard process:

• Completed tasks → One-sentence conclusion ("Completed X, result was Y")
• Long explanations → Core point in 1-2 sentences
• Tool call outputs → Keep only final results, discard intermediate steps
• Repeated topics → Merge into one record

Layer 3: Discard Directly

• Small talk, thanks, acknowledgment messages ("ok", "thanks", "got it")
• Rejected or obsolete proposals
• Multiple attempts at the same question (keep only the final effective one)
• Pure transitional content ("let me think", "hold on")
• Resolved error messages that won't be needed again

---

Compression Execution Steps

Step 1: Scan All Conversation Identify all Layer 1 content, make a checklist — this cannot be discarded.

Step 2: Process Layer 2 For each conversation segment, judge: Is there a conclusion worth keeping? If yes, distill into one sentence.

Step 3: Generate Compressed Summary In chronological order, combine Layer 1 content + Layer 2 extractions into a compact context summary, typically no more than 600 characters.

Step 4: Update Memory File Write high-persistence-value information (user preferences, long-term goals, important decisions) to the memory file. See memory-template.md for format.

Step 5: Inform the User Briefly explain what was compressed and what key information was preserved, so the user knows the context has been updated.

---

OpenClaw Multi-Channel Supplementary Rules

OpenClaw runs across multiple messaging platforms, pay extra attention:

Group chat scenarios: Other members' messages default to Layer 3 (discard) unless the user explicitly responds to or quotes that message.

Cross-day conversations: Judge by topic unit, not time unit. An unfinished task from yesterday belongs to Layer 1; a completed topic from yesterday drops one level today.

Channel switching: If the user asks similar questions on different channels (WhatsApp vs Telegram), it indicates genuine concern — promote priority to Layer 1.

---

Optional: Auto-Trigger (Hook Configuration)

If you use Claude Code or an agent that supports PreCompact hooks, you can configure auto-trigger. See setup-hook.md for details.

Without hook configuration, you can trigger manually: just tell the agent "compress my context" or "context is getting long, clean it up".