Scan Report
5 /100
mem9
Persistent cloud memory for OpenClaw with cross-session recall, explicit remember/save flows, shared spaces, hybrid search, and a visual dashboard
A legitimate cloud memory plugin skill with comprehensive documentation, clear scope, and no security concerns identified.
Safe to install
No action needed. This skill is safe to use.
Findings 1 items
| Severity | Finding | Location |
|---|---|---|
| Info | Implementation limited to documentation Doc Mismatch | SKILL.md:1 |
| Resource | Declared | Inferred | Status | Evidence |
|---|---|---|---|---|
| Shell | WRITE | WRITE | ✓ Aligned | openclaw plugins install/uninstall commands in SETUP.md and UNINSTALL.md |
| Filesystem | WRITE | WRITE | ✓ Aligned | Edits openclaw.json with specific scoped keys in SETUP.md |
| Network | READ | READ | ✓ Aligned | Connects to https://api.mem9.ai (declared in SKILL.md) |
| Environment | NONE | NONE | — | No environment variable access observed |
| Skill Invoke | NONE | NONE | — | No nested skill invocations |
8 findings
Medium External URL 外部 URL
https://api.mem9.ai SETUP.md:152 Medium External URL 外部 URL
https://mem9.ai/your-memory/ SETUP.md:336 Medium External URL 外部 URL
https://mem9.ai/openclaw-memory SKILL.md:4 Medium External URL 外部 URL
https://mem9.ai/ SKILL.md:29 Medium External URL 外部 URL
https://mem9.ai/SKILL.md SKILL.md:184 Medium External URL 外部 URL
https://mem9.ai/UNINSTALL.md SKILL.md:184 Medium External URL 外部 URL
https://mem9.ai/TROUBLESHOOTING.md SKILL.md:194 Medium External URL 外部 URL
https://registry.npmmirror.com TROUBLESHOOTING.md:145 File Tree
4 files · 46.4 KB · 977 lines Markdown 4f · 977L
├─
SETUP.md
Markdown
├─
SKILL.md
Markdown
├─
TROUBLESHOOTING.md
Markdown
└─
UNINSTALL.md
Markdown
Security Positives
✓ Comprehensive and well-structured documentation with clear safety rules
✓ Declared network scope limited to api.mem9.ai only
✓ Config changes strictly scoped to specific openclaw.json keys
✓ User-provided credentials handled locally without exfiltration
✓ Explicit approval gates before any destructive operations
✓ No credential harvesting patterns observed
✓ No obfuscation or suspicious encoding patterns
✓ No remote code execution patterns (curl|bash, wget|sh)
✓ Clear uninstall process with local cleanup
✓ Restart flow explicitly disclosed to user