@luis1213899

xia-anquan monitors three dimensions (capability, identity, knowledge) of openclaw agent state for persistent pollution attacks, credential leakage, and suspicious scripts. provides command-line audit and daemon modes with snapshot/alert logging.

skill-scoreboard tracks openclaw skill usage and assigns points based on invocation duration and complexity weight, supporting daily/weekly/monthly/yearly/all-time leaderboards with automated snapshots and workflow review reports.

codex-adapter-toolkit bridges codex responses api to multiple llm providers (deepseek, openai, gemini, ollama) via local http adapter with failover and circuit breaker patterns. primarily powershell/python based setup.

xia-zhua-zhua claims multi-modal content extraction, pdf export, and video download capabilities via command-line tools, but provides no actionable procedures, error handling, or input/output contracts beyond example invocations.

dream-selfimproving proposes nightly ai memory distillation via m-flow knowledge graphs and long-term rag, but lacks concrete actionable procedures, specific trigger phrases, and failure mode documentation. the skill reads as architectural vision rather than implementable workflow.

即梦AI生图与视频生成工具。通过火山引擎API调用即梦AI，支持文生图、图片生成、视频生成。内置50+提示词模板覆盖电商、社交、商业设计、视频等场景。

Create stunning, animation-rich HTML presentations from scratch or by converting PowerPoint files. Use when the user wants to build a presentation, convert a...