ClawHub
Back to skill

Security audit

damage-control

Security checks for vulnerabilities and agentic risk

Overview

This is a documented recovery playbook for agent crashes, context loss, and git conflicts, with some broad workflow guidance but no hidden execution or exfiltration behavior.

Install this if you want agents to follow structured recovery procedures for crashes, context overflow, merge conflicts, and inconsistent repo state. Because it can steer agents toward git stashes, commits, merge aborts, and task-note updates, keep it tied to explicit recovery or checkpoint situations and avoid broad automatic invocation in unrelated tasks.

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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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 (4)

Vague Triggers

Medium
Confidence
96% confidence
Finding
The trigger list includes broad, common terms like 'recovery', 'resilience', 'context', 'conflicts', and 'state', which can cause the skill to activate in unrelated situations. Unintended invocation of a recovery-oriented skill can disrupt normal workflows, bias the agent toward rollback/reconciliation actions, and introduce unnecessary handling of git state or task state when no failure exists.

Intent-Code Divergence

Medium
Confidence
93% confidence
Finding
The Overview states, "The skill does not prevent failures" and "defines what to do after one has already happened." Later, the Integration Pattern instructs orchestrators to invoke state-reconciliation at session boundaries regardless of whether a failure occurred, which directly contradicts the earlier stated intent of post-failure-only usage.

Skill Enumeration

Medium
Category
Agent Snooping
Content
- `../../risk-classification/modules/readiness-levels.md` - Level
  definitions
- `../SKILL.md` - Error recovery procedures
- `../../egregore/skills/quality-gate/SKILL.md` - Verification requirements
Confidence
80% confidence
Finding
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Description-Behavior Mismatch

Low
Confidence
79% confidence
Finding
The manifest describes the skill as recovering broken agent state after crashes, context overflow, and merge conflicts. However, the body also mandates a pre-execution "Risk Assessment Checklist" for Level 1+ tasks and recommends state-reconciliation at session boundaries even when no failure occurred, which extends beyond post-failure recovery into proactive operational process guidance.

Static analysis

No suspicious patterns detected.