Back to skill

Security audit

Reader

Security checks for vulnerabilities and agentic risk

Overview

This skill is a local document reader that reads user-specified text files and keeps a local history log, with no evidence of network access, credential use, deception, or destructive behavior.

Before installing, be aware that summaries, briefs, and comparisons can leave a local history record containing filenames or paths and timestamps. Avoid using it on highly sensitive documents if retaining that metadata locally is unacceptable.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (7)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
86% confidence
Finding

The skill advertises local file read/write behavior and persistent storage, but it does not declare any explicit tool scope or permissions boundaries. In an agent ecosystem, this can cause the skill to be invoked with broader-than-expected filesystem access, increasing the chance of unintended reads or writes to local data.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The description is very broad ('use whenever the user wants to read, summarize, extract key points, compare documents...') and could match many generic user prompts. Over-broad invocation criteria can cause the skill to activate unexpectedly on unrelated tasks, exposing local file handling and persistence features in contexts where the user did not intend them.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Low
Category
Not specified by scanner
Confidence
78% confidence
Finding

The manifest frames the skill as a local-first document reading and distillation engine, but this script also records comparison session history by writing file names and timestamps to storage. While still local, persistent tracking of user activity is additional behavior beyond simple read/compare processing and is not suggested by the description.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
86% confidence
Finding

The script writes comparison session metadata to persistent history without any user-facing notice or consent. Even though it does not store file contents, file paths and timestamps can reveal sensitive project names, usernames, directory structures, or document existence, which creates a local privacy exposure inconsistent with the tool's local-reading framing.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Low
Category
Not specified by scanner
Confidence
91% confidence
Finding

The manifest frames the skill as a local-first reading and distillation engine focused on processing documents and pasted text. This script goes beyond pure reading/distillation by recording session metadata via load_history/save_history, which is an additional persistence behavior not suggested by the description.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Low
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill is described as local-first reading/distillation with no external sync or credentials, but it still persists per-use metadata to storage. Even though the stored data appears limited to style, file path, and timestamp rather than document contents, it creates an undisclosed local activity log that may reveal sensitive filenames, user behavior, or document access patterns.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

The script writes usage history to storage without notifying the user at runtime or in the interface, which undermines user expectations for a privacy-preserving local reader. In this skill context, undisclosed persistence is more concerning because users may process sensitive local documents and assume no retained traces beyond the immediate summary output.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.