← Back to Skills Marketplace
255
Downloads
0
Stars
1
Active Installs
11
Versions
Install in OpenClaw
/install memory-trace
Description
Extract character personality from novels/scripts/anime, generate SoulPod packages for Memory-Inhabit dialogue. | 从小说、剧本、动漫素材中提取角色人格，生成 SoulPod 包。下游配合 Memory...
Usage Guidance
This skill appears to do what it claims (analyze text and build persona packages), but it also: 1) will perform web searches and download images/audio automatically; 2) will try to install another skill (multi-search-engine) if the agent lacks web-search capability; and 3) will write generated SoulPod files into a Memory-Inhabit/personas path (i.e., modify another skill's data). Before installing or enabling it, consider: - Review the included scripts (analyzer.py, forge.py, voice_catalog.json) locally to confirm there are no unwanted network calls or hidden endpoints. - Run the workflow manually in a sandbox or test environment first (do not allow automatic installs) to inspect downloaded assets and generated files. - Back up your Memory-Inhabit/personas directory before allowing automatic deployment. - Disable or require confirmation for automatic installation of other skills (multi-search-engine) or network downloads. - Ensure pdfplumber and any other dependencies are installed/managed by you rather than by the skill. - If you are concerned about copyright or sensitive content, restrict or vet the web sources the skill is allowed to pull from. If you want me to, I can: (a) list exact places in the scripts that perform filesystem/network operations, (b) suggest edits to require user confirmation before downloads/installs, or (c) produce a safe, sandboxed run plan you can follow.
Capability Analysis
Type: OpenClaw Skill
Name: memory-trace
Version: 1.0.10
The memory-trace skill bundle is designed to automate the extraction of character personalities from text and generate configuration packages (SoulPods) for roleplay. The Python scripts (analyzer.py and forge.py) perform legitimate text processing, regex-based character recognition, and file management. Notably, forge.py includes a sanitize_name function to mitigate path traversal risks when creating directories. While the SKILL.md instructions direct the AI agent to autonomously install dependencies and execute workflows, these actions are aligned with the stated purpose of the tool. No evidence of data exfiltration, malicious execution, or harmful prompt injection was found.
Capability Assessment
Purpose & Capability
Name/description align with the shipped code and sample outputs: analyzer.py and forge.py implement text analysis, personality inference, voice selection, and generation of SoulPod files (profile.json, system_prompts.txt, memories, assets). No unexpected credentials or unrelated binaries are requested.
Instruction Scope
SKILL.md instructs the agent to perform web search, download images/audio, save source material, run analyzer.py and forge.py, and then install the generated pod — and it explicitly says this runs '自动完成，无需等待' (automatic without confirmation). It also instructs auto-install of a separate skill (multi-search-engine) if network search capability is missing. Those instructions cause network I/O, file downloads, and automatic execution without user confirmation, which is broader scope than a simple local analysis tool and could lead to unexpected downloads or side effects.
Install Mechanism
There is no install spec (instruction-only), which lowers supply-chain risk. The package includes local Python scripts and a voice catalog; no external archive downloads or remote installers are present. A dependency (pdfplumber) is documented; the code will fail if not installed but does not itself perform remote installs. Still: SKILL.md's step to auto-install another skill from SkillHub is an installation action triggered at runtime (not via a locked install manifest).
Credentials
The skill requests no environment variables or credentials (good), but its runtime behavior writes output into a sibling project's Memory-Inhabit/personas directory (forge.py uses a relative SOULPOD_DIR pointing to ../Memory-Inhabit/personas). That means it will modify another skill's data/config paths, which is a cross-skill side effect not declared in metadata. Automatic downloading of images/audio from the web may also pull arbitrary external content into local storage.
Persistence & Privilege
always:false and default autonomous invocation are fine, but the instructions include automatic installation of a third-party skill (multi-search-engine) and automatic deployment into Memory-Inhabit/personas with no confirmation. Modifying other skills' directories and installing other skills at runtime increases blast radius and should be consented to explicitly by the user.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install memory-trace - After installation, invoke the skill by name or use
/memory-trace - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.10
Added 2 zhuangfy images, removed 5 legacy images and obsolete thunderstorm.txt | 新增庄fy图片2张，删除旧图5张及过时样本文件
v1.0.9
v1.0.9: 1.tts.py音色匹配按gender三维查表; 2.profile.json补全gender字段说明; 3.移除过时同步规则; 4.更新tts.py硬编码路径为新版repo结构
v1.0.8
v1.0.8: 音色目录统一管理 + gender 性别支持 + Edge 音色更新；工作流新增自动安装搜索技能 + 全流程自动执行无需等待
v1.0.7
v1.0.7: 新增音色推测功能，生成SoulPod时自动推断最优音色写入config.json
v1.0.6
v1.0.6: 添加官网主页链接
v1.0.5
- Expanded workflow: Added automated web search, media downloading (images and audio), and detailed data directory structure for the full character copying process.
- Enhanced documentation: Now describes the SoulPod directory, required output files, and profile.json fields.
- Included step-by-step instructions for both full automation and manual triggering methods.
- Listed synchronization rules for maintaining consistency across multiple SoulPod storage locations.
v1.0.4
补充叶修角色档案，author字段更新
v1.0.3
- Added output generation for a new sample ("夏以昼") including profile, system prompts, and raw memories files.
- Updated author information in SKILL.md.
- Internal script updates in scripts/forge.py to support new output structure.
v1.0.2
- Added _meta.json file for improved metadata support.
- Updated version number to 1.0.2 in SKILL.md.
- No feature changes or user-facing updates.
v1.0.1
- Added SKILL.md documentation, detailing the purpose, workflow, output, dependencies, and usage instructions.
- Provided a clear overview of how to extract and generate persona profiles (SoulPod) from creative sources.
- Specified setup requirements and simple usage commands for end users.
v1.0.0
Initial release of memory-trace.
- Extracts character personalities from novels, scripts, and anime materials.
- Generates a SoulPod package containing profile.json, system_prompts.txt, and memories.
- Supports personality cloning, personality analysis, and SoulPod creation.
- Simple trigger phrases and file-based workflows.
- Requires `pdfplumber` for processing PDF materials.
Metadata
Frequently Asked Questions
What is Trace?
Extract character personality from novels/scripts/anime, generate SoulPod packages for Memory-Inhabit dialogue. | 从小说、剧本、动漫素材中提取角色人格，生成 SoulPod 包。下游配合 Memory... It is an AI Agent Skill for Claude Code / OpenClaw, with 255 downloads so far.
How do I install Trace?
Run "/install memory-trace" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Trace free?
Yes, Trace is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Trace support?
Trace is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Trace?
It is built and maintained by Asuka (@evangeliona); the current version is v1.0.10.
More Skills