ClawHub
Back to skill

Security audit

prompt-architect-p

Security checks for vulnerabilities and agentic risk

Overview

This is a markdown-only prompt-building skill with some prompt-quality caveats but no hidden execution, persistence, credential access, or exfiltration behavior.

This skill appears safe to install for prompt drafting, but review generated prompts before reuse, especially prompts that ask models to show full reasoning or handle sensitive documents. Consider changing such prompts to request a concise rationale instead of hidden/internal reasoning.

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
  • System Prompt LeakageDirect Leakage, Indirect Extraction, Tool-Based Exfiltration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Direct Prompt Extraction

High
Category
System Prompt Leakage
Confidence
85% confidence
Finding

Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.

Content

Scanner excerpt · SKILL.md (reported line 60)May include surrounding context.

md
- Appropriate framework from `references/frameworks.md`
- Quality criteria from `references/quality-criteria.md`

**Output rules:**
- Deliver inside a **code block** for easy copying
- Include a brief note explaining which framework was used and why
- If the prompt is complex, add inline comments

Direct Prompt Extraction

High
Category
System Prompt Leakage
Confidence
85% confidence
Finding

Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.

Content

Scanner excerpt · references/frameworks.md (reported line 13)May include surrounding context.

md
7. [Self-Consistency](#self-consistency)
8. [Prompt Chaining](#prompt-chaining)
9. [Meta-Prompting](#meta-prompting)
10. [Structured Output Prompting](#structured-output-prompting)
11. [Constitutional / Guardrail Prompting](#constitutional--guardrail-prompting)
12. [Multimodal Prompting](#multimodal-prompting)

Direct Prompt Extraction

High
Category
System Prompt Leakage
Confidence
85% confidence
Finding

Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.

Content

Scanner excerpt · references/frameworks.md (reported line 228)May include surrounding context.

md
7. [Self-Consistency](#self-consistency)
8. [Prompt Chaining](#prompt-chaining)
9. [Meta-Prompting](#meta-prompting)
10. [Structured Output Prompting](#structured-output-prompting)
11. [Constitutional / Guardrail Prompting](#constitutional--guardrail-prompting)
12. [Multimodal Prompting](#multimodal-prompting)

Direct Prompt Extraction

High
Category
System Prompt Leakage
Confidence
85% confidence
Finding

Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.

Content

Scanner excerpt · references/frameworks.md (reported line 108)May include surrounding context.

  • [Quality 1 of the persona]
  • [Quality 2 of the persona]

[Output instructions]

text

**Tips:**

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The instructions require asking exactly whether the final prompt should be in English or Arabic, which restricts output language to two options regardless of user preference. This is a natural-language policy concern because it imposes a locale/language constraint without offering a general choice or documenting why the limitation is necessary.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The Chain-of-Thought section explicitly instructs the model to 'Show your reasoning before giving the final answer,' which encourages disclosure of intermediate reasoning. In a prompt-engineering skill, this can normalize requesting hidden reasoning and may lead downstream users to elicit sensitive internal analysis, policy text, or unnecessary verbose traces without any warning about safety, privacy, or prompt-leakage implications.

Content

No source excerpt is available for this finding.