MCP App StoreMCP App Store
from claw-compactor2,322
Six-layer token-compression toolkit for OpenClaw workspaces (rule engines, dictionary encoding, observation compression, RLE, protocol optimizations, plus Engra
Claw Compactor provides a multi-layer pipeline to compress workspace content and reduce LLM token costs. It combines deterministic rule-based compression (dedup, dictionary encoding, RLE, tokenizer optimizations) with an optional Engram LLM-driven Observational Memory layer for semantic summarization and long-term reflections. Includes CLIs to run benchmarks, full pipelines, and targeted layer runs.
Run at session start to shrink injected system context, before expensive LLM calls to lower cost, or as a regular maintenance cron to compress accumulated memory files. Use benchmark mode to estimate savings; run full pipeline when savings justify writes. Useful for teams managing large agent workspace histories.
Agents with workspace access and ability to run Python/CLI (OpenClaw, CLI-capable automation agents). Engram layer requires an LLM-compatible endpoint or API key for Anthropic/OpenAI-compatible services.
Claw Compactor is a 6-layer token-compression toolkit for OpenClaw workspaces, combining deterministic rule engines with an LLM-driven Engram observational memory system. The SKILL.md is thorough and well-structured with clear commands, configuration, and integration docs. However, every single script fails at import time because the core `claw_compactor` Python package is not included in the skill bundle — only the scripts/ directory was fetched. This makes the skill entirely non-functional in its current discovered form. The code itself is well-written with good docstrings and error handling, but cannot be evaluated beyond static analysis.
claw_compactor (Python package — not installable via pip, expected at lib/claw_compactor/)The skill has a clear separation between scripts/ and an expected lib/claw_compactor/ package, but only scripts/ was fetched for audit. The missing package makes all scripts non-functional. No security concerns — no hardcoded secrets, no remote code execution patterns, no destructive commands. The skill is well-intentioned and would be genuinely useful if the dependency were bundled or installable. Minor shell quoting concern in engram-auto.sh but not exploitable in normal use.
Extract Circuit Netlist Edge Features
Extract structured edge features from a bipartite circuit netlist (NetworkX MultiGraph), normalizing device/net ordering, mapping terminal colors, and detecting
Retriever (Membrane CLI)
Integrate Retriever search and discovery via the Membrane CLI — manage connections and run queries across Retriever data sources.
Yielding Bear — Unified LLM Routing
A unified LLM routing API that automatically routes requests to the cheapest capable model across providers to reduce inference cost by ~60-80%.
Caveman Compress
Compress natural-language memory files into a compact 'caveman' format to save LLM tokens while preserving code, links, and structure.
Markdrop
Convert PDFs to structured Markdown or interactive HTML and generate AI-powered descriptions for images and tables using multiple LLM providers.
FastAPI Project Templates
Creates production-ready FastAPI project scaffolds with async patterns, DI, middleware, and testing best practices for high-performance APIs.
Unitree Robot Controller
Control and command Unitree robots (GO1/GO2/G1/H1) via OpenClaw: initialization, basic motion commands, and sensor integrations.
Alpha Forge Pre-Ship Quality Gates
Pre-merge quality gates for PRs that validate RNG determinism, forked URLs, runtime parameter ranges, and manifest synchronization to reduce review cycles.