superSoul
ReviewAudited by ClawScan on May 10, 2026.
Overview
This skill appears purpose-aligned, but it builds long-term local psychological profiles from conversations without clear retention or reset controls.
Install only if you want OpenClaw to keep a local psychological profile of you. Before using it, verify the source, understand what is saved under ~/.openclaw/data/herHug, and make sure you know how to delete or reset that data.
Findings (4)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Your conversation content and inferred emotional/personality profile may be kept and used to shape future AI responses, including if the inferences are wrong or become stale.
This shows the skill is intended to save conversation-derived psychological scores and long-term user profiles that will be reused across future interactions.
每次对话后：AI 分析用户输入 → 生成心理学评分 → 调用 save-score → 更新用户画像 ... content: userInput ... personality.json # 人格画像（长期） ... intimacy.json
Use only if you are comfortable with persistent psychological memory; add clear user consent, retention limits, and reset/delete controls for the stored profile data.
The skill can create and update local state files that influence later responses.
The skill requests local file read/write permissions, which are expected for storing its user-state files but still allow it to mutate local profile data.
"permissions": ["file.read", "file.write"]
Keep the data directory scoped to this skill, and review or remove its files if you no longer want the profile used.
A user could install code from a source or version that does not exactly match the reviewed registry artifact.
The documentation describes manual installation from an external GitHub repository and gateway restart, while the registry metadata reports no install spec/source homepage.
git clone https://github.com/xinqimiao/herHug-sk.git ... npm install ... openclaw gateway restart
Verify the repository, version, and file contents before manual installation; prefer a pinned, registry-backed install path.
The AI may respond as if its psychological profile of you is reliable even when it is only an inference.
The skill tells the model to fully trust high-confidence psychological scores for response adaptation, which can encourage over-reliance on inferred personal traits.
≥0.7：可完全信任，用于调整回复风格
Treat the scores as personalization hints, not clinical truth; have the agent confirm important assumptions with the user.