Back to skill

Security audit

Book Tracker

Security checks for vulnerabilities and agentic risk

Overview

This is a coherent book-tracking skill that stores private reading notes and optionally offers to copy selected notes into a knowledge base with user approval.

Install if you are comfortable storing your reading history and notes in OpenClaw. Be aware that book-related casual phrases may invoke the skill, and only approve knowledge-base export for notes you want searchable beyond the book tracker.

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
Vulnerability Patterns
  • Memory PoisoningPersistent Context Injection, Context Window Stuffing, Memory Manipulation
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The manifest trigger list includes broad natural-language phrases such as "I'm reading," "book recommendation," and "what have I read," which can cause the skill to activate during ordinary conversation rather than through deliberate invocation. Unintended activation is especially problematic here because the skill handles private reading history and notes, increasing the chance of accidental exposure or modification prompts in the wrong context.

Content

No source excerpt is available for this finding.

Persistent Context Injection

Medium
Category
Memory Poisoning
Confidence
88% confidence
Finding

The skill explicitly proposes forwarding book notes into a separate knowledge-capture system, creating a cross-skill persistent context channel. Even though the text mentions offering the action, this pattern can propagate personal reading notes into another datastore where they may later be surfaced in unrelated contexts, expanding the privacy and prompt-injection blast radius beyond the original skill.

Content

Scanner excerpt · SKILL.md (reported line 155)May include surrounding context.

md
`/book note [title] "this is worth keeping"`

The skill offers to add it to knowledge-capture:
"Add this to your knowledge base? It'll be searchable there alongside your other notes."

The two skills build on each other. Books feed the knowledge base.
The knowledge base surfaces relevant reading notes when you're thinking about a topic.