ClawHub
Back to skill

Security audit

Agent Ops Hardening

Security checks for vulnerabilities and agentic risk

Overview

This skill provides disclosed agent operations-hardening guidance and a read-only audit script, with persistent memory guidance that fits its stated purpose.

Review the persistent memory and heartbeat recommendations before adopting them, since they can shape future agent behavior. The bundled audit script is read-only, but it references expected workspace files and a rick-vault path, so run it only in workspaces where those checks make sense.

A.I.G

Vulnerability Patterns
  • 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
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Memory PoisoningPersistent Context Injection, Context Window Stuffing, Memory Manipulation
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
85% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · SKILL.md (reported line 40)May include surrounding context.

NO

rm myfile.txt rm -rf ./important-directory

text

- Any file deletion should use `trash` or `mv` to archive unless explicitly intended as permanent

Persistent Context Injection

Medium
Category
Memory Poisoning
Confidence
80% confidence
Finding

Skill injects content designed to persist in agent memory or context across interactions. Persistent injection can alter agent behavior long after the initial interaction.

Content

Scanner excerpt · SKILL.md (reported line 160)May include surrounding context.

md
1. Audit MEMORY.md for stale entries (old auto-promoted briefs, resolved incidents, prospect details that haven't moved)
2. Move stale content to MEMORY-COLD.md (never delete)
3. Compress verbose sections into single-line rules
4. Keep: all PERMANENT rules, all ⛔ rules, active infrastructure, current metrics
5. Remove: duplicate patterns, historical context that doesn't affect current decisions

**Target:** Under 200 lines hot, unlimited cold. Nothing is ever deleted — it just moves tiers.