OpenClaw 集中配置管理系统

Security checks across static analysis, malware telemetry, and agentic risk

Overview

This appears to be a legitimate OpenClaw configuration-template skill, but it needs review because it recommends unverified external installers and includes credential-permission and privacy-scope issues.

Install only if you are comfortable making OpenClaw-wide configuration changes. Before using it, fix the secret-file permission guidance, inspect any scripts it asks you to run, avoid blind `curl | bash` installs, verify ClawRouter's data handling before funding or routing prompts through it, and keep backups of your OpenClaw config and memory files.

Static analysis

No static analysis findings were reported for this release.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Risk analysis

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.

What this means

API keys or app secrets stored in OpenClaw config files could become readable by other local users or processes on the machine.

Why it was flagged

The skill warns that some config files contain secrets, but also documents blanket/world-readable `644` permissions and a loader snippet that resets loaded config files to `644`. The provided templates store Feishu, ClawRouter, and SiliconFlow-style API secrets in JSON config files.

Skill content
chmod 644 ~/.openclaw/config/*.json ... chmod 600 ~/.openclaw/config/channels/feishu.json  # 密钥文件更严格 ... chmod 644 "$file" 2>/dev/null
Recommendation

Do not use the unconditional `chmod 644` behavior for secret-bearing files. Keep files containing tokens or app secrets at `600`, consider environment variables or a secrets manager, and verify permissions after running any setup script.

What this means

If the remote script, container image, or upstream source is compromised or changes unexpectedly, installing it could execute arbitrary code on the user's machine.

Why it was flagged

The recommended ClawRouter install path executes a remote shell script directly and also documents an unpinned Docker `latest` image, without checksums, signatures, version pins, or review steps.

Skill content
curl -fsSL https://blockrun.ai/ClawRouter-update | bash ... docker run ... blockrun/clawrouter:latest
Recommendation

Inspect the installer before running it, prefer pinned releases or checksummed artifacts, avoid `latest` tags, and run the service with least privilege.

What this means

A user may assume prompts and cached content never leave their machine, even though model routing may send request content to external providers.

Why it was flagged

The template makes a strong privacy claim that data stays local/does not leave, while the same configuration routes requests to third-party model providers. That claim is not sufficiently bounded or explained in the artifacts.

Skill content
"fallback_chain": ["anthropic/claude-sonnet-4", "zhipu/glm-4", "deepseek/deepseek-chat", "openai/gpt-4o-mini"] ... "ClawRouter 本地运行,数据不出境"
Recommendation

Verify ClawRouter's actual data flow, provider routing, logging, caching, retention, and encryption behavior before sending sensitive prompts through it.

What this means

These commands can change global OpenClaw behavior and active agent state if run on a real installation.

Why it was flagged

The setup flow instructs users to run scripts that generate the main configuration, synchronize memory/persona state, and force-restart the OpenClaw gateway.

Skill content
~/.openclaw/scripts/generate-main-config.sh ... ~/.openclaw/scripts/update-soul.sh ... openclaw gateway restart --force
Recommendation

Back up `~/.openclaw/openclaw.json` and relevant memory/persona files, inspect any scripts before running them, and avoid force restarts until the generated config has been reviewed.

What this means

Private context may be stored long-term in memory files and, if memorySearch is enabled, sent to an external embedding API for indexing/search.

Why it was flagged

The memory templates enable persistent automatic memory writes and semantic memory search using a remote embedding provider.

Skill content
AI 自动将关键信息写入 memory/YYYY-MM-DD.md ... "baseUrl": "https://api.siliconflow.cn/v1", "apiKey": "{{SiliconFlow API Key}}"
Recommendation

Define which memory paths may be indexed, exclude secrets and highly sensitive notes, document retention/deletion rules, and only enable remote memory search after confirming the provider's privacy terms.