ClawHub
Back to skill

Security audit

Coordinator Agent

Security checks for vulnerabilities and agentic risk

Overview

The skill is openly a fleet-monitoring coordinator, but it asks for broad workspace reads, external message delivery, optional credential copying, and optional cron retriggering that users should review carefully.

Install only if you are comfortable with a coordinator reading the specific workspaces you configure and sending summarized content to external chat platforms. Use dedicated scoped credentials, do not copy another agent's auth profile, exclude secrets and sensitive workspaces, use private messaging destinations, and leave self-healing disabled unless each retriggerable cron job is idempotent and approved.

A.I.G

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Findings (4)

T05 · Unauthorized Access and Privilege Escalation

Error
Location
SKILL.md:22
Finding

Unrestricted workspace aggregation may expose sensitive data to external messaging platforms

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:85
Finding

Authentication profiles can be copied from a more privileged agent

Content
View full analysis
Remediation
View remediation

T05 · Unauthorized Access and Privilege Escalation

Error
Location
SKILL.md:44
Finding

Self-healing functionality can retrigger scheduled jobs without an enforceable safety boundary

Content
View full analysis
Remediation
View remediation

T08 · Insecure Dependencies

Warning
Location
SKILL.md:52
Finding

Installation command retrieves an unpinned third-party package

Content
View full analysis
Remediation
View remediation

SkillSpector

Vulnerability Patterns
  • Rogue AgentSelf-Modification, Session Persistence
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · SKILL.md (reported line 27)May include surrounding context.

md
- **Workspace access:** The coordinator scans every workspace you configure. Restrict the workspace list to only what needs monitoring. Exclude any workspace containing credentials, secrets, or sensitive data.
- **Outbound messaging:** Briefing content is sent to Telegram (always) and optionally Discord/Slack. Review what gets included in the output format before enabling delivery to external platforms.
- **Auth scoping:** Use dedicated, scoped API credentials for the coordinator agent. Do not reuse auth profiles from other agents. The coordinator only needs read access to workspaces and write access to configured messaging channels.
- **Self-healing:** Automatically retriggering missed cron jobs can cause duplicate execution. Only enable this if your cron jobs are idempotent (safe to run twice). Disable self-healing by removing the retrigger rule from your SOUL.md if jobs have side effects.

## How It Works

Static analysis

No suspicious patterns detected.