ClawHub
Back to skill

Security audit

Buku

Security checks for vulnerabilities and agentic risk

Overview

This bookmark-management skill is purpose-aligned, but it normalizes deletion commands that bypass confirmation and can remove ranges of saved bookmarks.

Install only if you are comfortable with an agent managing your local buku database. Before any delete, range delete, tag deletion, import, or export, require the agent to show the matching bookmarks and get your explicit confirmation of the exact index, range, or file path.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (4)

Vague Triggers

Medium
Confidence
95% confidence
Finding
The trigger phrases are broad enough to match common conversational language about links and bookmarks, which can cause the skill to activate when the user did not explicitly intend to manage bookmarks. In an agent setting, overbroad activation increases the chance of unintended tool use and can lead to accidental bookmark creation, modification, import/export, or deletion workflows.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The delete section documents destructive commands, including range deletion, without any user-facing warning about permanent data loss or a requirement to confirm intent. In a bookmark-management skill, this creates a realistic risk of accidental mass deletion if the agent misparses an index or the user request is ambiguous.

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
### Delete
```bash
# By index (use --tacit to skip confirmation)
buku --nostdin --tacit -d INDEX

# Range
Confidence
95% confidence
Finding
Using `--tacit` to skip confirmation for deletion enables autonomous destructive action without a user verification checkpoint. In this skill, that is especially risky because the same document also promotes broad triggering and supports deletion by single index or range, making accidental or unauthorized bookmark loss more likely.

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
- **Always use `--nostdin`** as first arg to prevent waiting for input
- **Always use `--np`** for search/print to skip interactive prompt
- **Use `--tacit`** for delete operations to skip confirmation
- **Use `-j`** for JSON output when parsing results programmatically
- DB location: `~/.local/share/buku/bookmarks.db`
Confidence
94% confidence
Finding
The guidance to use `--tacit` for delete operations institutionalizes confirmation bypass as normal behavior for the skill. This removes a key safety barrier for destructive actions and increases the chance that an autonomous agent will delete bookmark data without adequate user awareness or validation.

Static analysis

No suspicious patterns detected.