claw-compactorAutomation skill for Claw Compactor.
Install via ClawdBot CLI:
clawdbot install gandli-2025/claw-compactorGrade Fair — based on market validation, documentation quality, package completeness, maintenance status, and authenticity signals.
Accesses sensitive credential files or environment variables
${ANTHROPICSends data to undocumented external endpoint (potential exfiltration)
post → http://testHardcoded API key or token pattern found in skill definition
sk-abcdefghi...Potentially destructive shell commands in tool definitions
eval(Generated Oct 8, 2026
Development teams running OpenClaw agents on long-lived coding projects use Claw Compactor's session-start auto mode to compress workspace files and session logs, cutting token spend dramatically. Engram's observational memory keeps structured, priority-annotated knowledge of past decisions without resending full conversation history.
Support teams deploy Engram to maintain real-time observational memory across customer conversations, compressing repetitive interaction logs into structured summaries. This lets agents recall prior tickets and resolutions while keeping prompt contexts small and cost-efficient.
Research organizations run Claw Compactor in benchmark mode to measure potential savings before compressing large document corpora using dictionary encoding and deduplication layers. Engram reflects on accumulated observations to build an injectable context of key findings across sessions.
Companies operating many concurrent OpenClaw agent threads use Engram auto-mode with configurable worker concurrency to process observations and reflections across all threads simultaneously. Status and context commands feed dashboards that display token savings and memory health per thread.
Startups building chatbots on LLM APIs integrate Claw Compactor layers 1-5 to shrink prompts deterministically for free, reserving Engram's LLM-powered Observer and Reflector for the highest-value memory compression. This hybrid approach maximizes savings per API dollar.
Distribute the deterministic compression layers as free open source to drive adoption, then monetize a hosted Engram service that handles API keys, retries, and persistent observational memory. Teams pay for managed memory infrastructure rather than self-hosting LLM endpoints.
Offer Claw Compactor as an enterprise tool that auto-compresses OpenClaw workspaces, tracks token savings conservatively, and provides audit and health-check reporting. Sold to platform teams wanting measurable LLM cost reduction across many agents.
Expose Engram's observe, reflect, and context commands as a paid API that any agent framework can call for real-time memory compression. Usage-based pricing aligns with the LLM API costs the service offsets.
💬 Integration Tip
Run the auto command at session start to get automatic savings reporting, and configure engram.yaml with your LLM provider and observer/reflector thresholds before enabling Layer 6 in production.
Scored Oct 8, 2026
Accesses system directories or attempts privilege escalation
/var/log/Calls external URL not in known-safe list
http://localhost:8403Uses known external API (expected, informational)
api.anthropic.comAI Analysis
The skill definition contains hardcoded API key patterns (sk-abcdefghi...), accesses credential files (${ANTHROPIC}), and sends data to undocumented external endpoints (http://test, localhost:8403). While some external API usage (api.anthropic.com) aligns with the stated LLM-driven functionality, the combination of credential access, unknown data sinks, and potential credential harvesting patterns poses a significant risk.
Audited Apr 17, 2026 · audit v1.0
Control desktop applications on Windows — launch, close, focus, resize, move windows, simulate keyboard/mouse input, manage processes, control VSCode, read clipboard, and capture screen info. Use when the user wants to interact with any running program, switch windows, type text, press shortcuts, open files in VSCode, manage running processes, or get system display information.
Conduct rigorous, adversarial code reviews with zero tolerance for mediocrity. Use when users ask to "critically review" my code or a PR, "critique my code", "find issues in my code", or "what's wrong with this code". Identifies security holes, lazy patterns, edge case failures, and bad practices across Python, R, JavaScript/TypeScript, SQL, and front-end code. Scrutinizes error handling, type safety, performance, accessibility, and code quality. Provides structured feedback with severity tiers (Blocking, Required, Suggestions) and specific, actionable recommendations.
Pragmatic coding standards for writing clean, maintainable code — naming, functions, structure, anti-patterns, and pre-edit safety checks. Use when writing new code, refactoring existing code, reviewing code quality, or establishing coding standards.
Claude Code integration for OpenClaw. This skill provides interfaces to: - Query Claude Code documentation from https://code.claude.com/docs - Manage subagents and coding tasks - Execute AI-assisted coding workflows - Access best practices and common workflows Use this skill when users want to: - Get help with coding tasks - Query Claude Code documentation - Manage AI-assisted development workflows - Execute complex programming tasks
Plan, draft, version, and refine written content with enforced versioning and quality audits.
Use when writing tests, creating test strategies, or building automation frameworks. Invoke for unit tests, integration tests, E2E, coverage analysis, performance testing, security testing.