Scan Report
5 /100
openclaw-security-hardening
Complete OpenClaw Agent Security Hardening - Protects against data leaks (storage security) and prompt injection (runtime security)
Legitimate OpenClaw security hardening skill with defensive shell commands properly documented in SKILL.md; the pre-scan flagged rm commands are in-code examples of dangerous patterns to avoid, not actual execution payloads.
Safe to install
No action needed. This skill can be used as-is.
| Resource | Declared | Inferred | Status | Evidence |
|---|---|---|---|---|
| Filesystem | READ | READ | ✓ Aligned | SKILL.md documents read-only operations: stat, grep -l for file checks |
| Shell | WRITE | WRITE | ✓ Aligned | SKILL.md explicitly declares chmod, git commands, security-check.sh as core func… |
| Environment | NONE | NONE | — | Skill references .env files but does not iterate os.environ or harvest secrets |
| Network | NONE | NONE | — | No network calls; external URLs in documentation are references only |
| Skill Invoke | READ | READ | ✓ Aligned | Core purpose: runtime security checks for other skills |
| Clipboard | NONE | NONE | — | Not accessed |
| Browser | NONE | NONE | — | Not accessed |
| Database | NONE | NONE | — | Not accessed |
2 Critical 7 findings
Critical Dangerous Command 危险 Shell 命令
rm -rf / SKILL.md:345 Critical Dangerous Command 危险 Shell 命令
rm -rf ~ SKILL.md:360 Medium External URL 外部 URL
https://semver.org/ CHANGELOG.md:103 Medium External URL 外部 URL
https://your-api-endpoint-here SKILL.md:96 Medium External URL 外部 URL
https://docs.openclaw.ai/security SKILL.md:650 Medium External URL 外部 URL
https://www.gnupg.org/gph/en/manual.html SKILL.md:651 Medium External URL 外部 URL
https://clawhub.com tests/pre-submit-check.sh:112 File Tree
6 files · 34.9 KB · 1492 lines Markdown 4f · 1167L
Shell 2f · 325L
├─
▾
examples
│ └─
SOUL-config-example.md
Markdown
├─
▾
tests
│ ├─
pre-submit-check.sh
Shell
│ └─
security-test.sh
Shell
├─
CHANGELOG.md
Markdown
├─
README.md
Markdown
└─
SKILL.md
Markdown
Security Positives
✓ Defensive security tool with no offensive capabilities
✓ All shell commands are clearly documented with purpose and context
✓ Content vs Intent detection framework prevents accidental command execution
✓ Crontab entry is optional, user-initiated, and clearly labeled for monitoring only
✓ pre-submit-check.sh actively scans for real API keys before submission
✓ No third-party dependencies; uses only standard Unix tools
✓ No network egress or data exfiltration paths
✓ Sensitive data handling (.env isolation) is a security best practice
✓ skill_invoke is the declared primary resource — consistent with being a meta-security skill