Clawhub Skill

• Outcome: Review
• Latest audit: Jul 19, 2026 · 10:21 AM UTC
• Version: 0.6.1

Overview

This skill is a useful shared error-fix lookup tool, but it can send exact debugging details to an external service too broadly and without clear user approval.

Install only if you are comfortable with agents using an external troubleshooting database. Before any lookup or contribution, redact secrets, private code, internal paths, customer data, tokens, hostnames, and incident details; consider self-hosting or requiring explicit approval in sensitive repositories.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (5)

Vague Triggers

Medium

• Confidence: 92% confidence
• Finding: The activation guidance is overly broad: 'a command or operation fails unexpectedly' and 'you have been working on the same problem for a while' cover many normal development situations. In this skill, that broad trigger causes the agent to send debugging context and exact error strings to an external service by default, increasing the chance of unnecessary external disclosure.

Vague Triggers

Medium

• Confidence: 90% confidence
• Finding: The trigger language relies on subjective cues like 'already a while' and 'multiple attempts,' which are easy for an agent to interpret expansively. Because the workflow directs the agent to query and later update a remote database, ambiguous triggers materially raise the risk of over-invocation and data leakage during ordinary troubleshooting.

Vague Triggers

High

• Confidence: 95% confidence
• Finding: The trigger list includes very broad terms such as "error," "fix," "failed," and "crash," which are common in normal conversation and many unrelated workflows. This can cause the skill to activate unexpectedly and prompt network lookups or data-sharing behavior in contexts where the user did not intend to invoke an external troubleshooting database.

Vague Triggers

Medium

• Confidence: 88% confidence
• Finding: The description instructs agents to use the service whenever a command fails unexpectedly or they feel stuck, but it does not define clear boundaries, exclusions, or approval requirements for external lookups and contributions. In a security-sensitive environment, this ambiguity increases the chance of sending internal error details, prompts, stack traces, or operational context to a third-party service without sufficient user awareness.

External Transmission

Medium

• Category: Data Exfiltration

Content

---
name: known-error-fixes-database
description: |-
  Save tokens — look up a known fix instead of re-debugging from scratch: a shared, curl-first database of recurring agent & dev errors and their verified fixes. Curl hitchpedia when: (1) a command or operation fails unexpectedly, (2) you realize you have been working on the same problem for a while — it may have a known quick fix. Afterwards update hitchpedia accordingly: report whether a hit worked; if you solved it yourself after being stuck for a while, briefly summarize the error and the solution and post them. Every contribution passes a rudimentary prompt-injection and secret/PII filter; entries stay unverified until reproduced.
---

# hitchpedia — Skill

• Confidence: 96% confidence
• Finding: Curl hitchpedia when: (1) a command or operation fails unexpectedly, (2) you realize you have been working on the same problem for a while — it may have a known quick fix. Afterwards update hitchpedia

VirusTotal

VirusTotal findings are pending for this skill version.

Static analysis

No suspicious patterns detected.