by 安全扫描
OpenClaw
可疑
medium confidenceThe skill's markdown instructions mostly match its stated purpose (local Markdown-based personality/memory), but there are inconsistencies and implicit behaviors (network/git operations, following external PROTOCOL.md steps) that need review before trusting it.
评估建议
This skill appears to implement what it claims (local Markdown 'soul' files) but it asks the agent to clone and pull a GitHub repository and to follow instructions from a protocol file that is not included in the skill package. Before installing or letting the agent run these steps: 1) Manually inspect the GitHub repo (https://github.com/LucioLiu/relic) and PROTOCOL.md to verify there are no scripts or commands you don't trust. 2) Note that the skill assumes git and internet access — if you don'...详细分析 ▾
ℹ 用途与能力
The skill claims to provide cross-agent personality/memory via local Markdown files and the SKILL.md instructions align with that goal (read/write ~/relic/brain/*). However the metadata declares no required binaries or install steps while the runtime instructions assume git or a zip download and network access. The missing declaration of those dependencies is an inconsistency.
⚠ 指令范围
Instructions direct the agent to clone/pull a GitHub repo and to 'Read PROTOCOL.md Section 0 and follow all steps.' Those external protocol steps are not included here and could require additional actions beyond the stated scope (file system modifications, config changes, or running scripts). The skill says 'NEVER access files outside ~/relic/brain/', but following external PROTOCOL.md blindly could expand the agent's actions. The agent is also told to perform network checks against raw.githubusercontent.com, which is outside purely local file operations.
ℹ 安装机制
There is no formal install spec in the registry, but SKILL.md instructs a git clone from github.com or downloading the repository zip. Fetching from GitHub is a standard source (lower risk than an arbitrary server), but this will write code to ~/relic and the repo contents are not included in the skill package for review. The absence of an explicit install declaration and the implicit reliance on git/network are notable.
ℹ 凭证需求
The skill declares no credentials or env vars and does not request secrets in the SKILL.md. That matches most of the instructions. However the protocol references (PROTOCOL.md) and optional 'neural anchor' auto-load (Section 9) could require additional configuration or permissions not shown here. Also the runtime behavior assumes network access for version checks and git pull (no credentials, but network access is required).
ℹ 持久化与权限
The skill metadata does not request permanent presence (always:false). The SKILL.md offers an optional auto-load setup (neural anchor) that could increase persistence if the user agrees. Because the exact steps for making Relic auto-load are in PROTOCOL.md (not included), that could change the persistence/privilege footprint and should be reviewed before enabling.
安全有层次，运行前请审查代码。
运行时依赖
无特殊依赖
版本
latestv0.1.42026/4/16
Added automatic version check (Step 0): agent checks GitHub for updates on every session, asks user before updating. Never force-updates.
● Pending
安装命令
点击复制官方npx clawhub@latest install relic-soul-chip
镜像加速npx clawhub@latest install relic-soul-chip --registry https://cn.longxiaskill.com