🎨 Dashboard: Visual HTML report with security score
Agent Instructions
Everything below is for the agent executing this skill.
When to Use
Trigger this skill when:
User requests security checkup/audit
Auto-trigger: Installing skills, creating/modifying agents, adding/modifying cron jobs
Periodic review (recommended: weekly)
Expertise Levels
Level
Identifier
Style
Beginner
1, beginner
Analogies, simple explanations, no jargon
Intermediate
2, intermediate
Technical details, config examples
Expert
3, expert
Attack vectors, edge cases, CVE references
Execution Flow
Step 1: Ask Level (before running anything)
Present options in user's language. Example (English):
What level of technical detail do you prefer?
1. 🌱 Beginner - I'll explain simply with analogies
2. 💻 Intermediate - Technical details and config examples
3. 🔐 Expert - Include attack vectors and edge cases
📌 All levels run the same checks—only explanation depth varies.
The script returns JSON with url (LAN IP) and local_url (localhost).
Use the url field (not localhost) when telling the user — they may access from another device.
Report what you can check and note Windows-specific limitations.
Minimal Environments (Docker, DSM)
Some environments lack tools. Check output and supplement:
Missing Tool
Fallback
curl
wget -qO-
ss
netstat
ip
ifconfig or /sbin/ip
pgrep
ps aux | grep
Agent Decision Flow
1. Run full_audit.sh
2. Check "failed_checks" in output
3. For each failed check:
a. Identify platform from os/os_variant
b. Run platform-specific fallback command
c. Incorporate results into report
4. Note any checks that couldn't complete
🔴 Critical - Immediate action required. Active exploitation possible.
🟠 High - Significant risk. Should fix soon.
🟡 Medium - Notable concern. Plan to address.
🟢 Low - Minor issue or best practice recommendation.
⚪ Info - Not a risk, but worth noting.
Risk Matrix
Tool Permissions
Minimal Full
┌──────────┬──────────┐
Exposure │ 🟢 │ 🟡 │
Low │ Safe │ Caution │
├──────────┼──────────┤
│ 🟡 │ 🔴 │
High │ Caution │ Critical │
└──────────┴──────────┘
Exposure = Who can talk to the bot (DM policy, group access, public channels)
Tool Permissions = What the bot can do (exec, file access, messaging, browser)
Context-Aware Exceptions
Don't just pattern match. Consider context:
Context
Adjustment
Private channel, 2-3 trusted members
Lower risk even with exec
VPN/Tailscale only access
Network exposure less critical
Self-hosted, single user
Session isolation less important
Containerized environment
Privilege escalation less severe
Always ask about environment if unclear.
Applying Fixes
CRITICAL RULES:
Never auto-apply fixes. Always show suggestions first.
Warn about functional impact. If a fix might break something, say so.
Get explicit user confirmation before any config changes.
Example flow:
Agent: "Changing this setting will disable exec in #dev channel.
If you're using code execution there, it will stop working.
Apply this fix?"
User: "yes"
Agent: [apply fix via gateway config.patch]
Language Rules
Internal processing: Always English
Thinking/reasoning: Always English
Final user-facing report: Match user's language
Technical terms: Keep in English (exec, cron, gateway, etc.)
Auto-Review Triggers
Invoke automatically when:
Skill installation: clawhub install <skill> or manual addition
Agent creation/modification: New agent or tool changes
Cron job creation/modification: New or modified scheduled tasks
For auto-reviews, focus only on changed component unless full audit requested.
