ClawHub
Back to skill

Security audit

Interview Architect

Security checks for vulnerabilities and agentic risk

Overview

This is a documentation-only hiring interview framework with no hidden execution, credential access, persistence, or data exfiltration behavior found.

Use this as a template generator, and avoid providing more candidate personal data than necessary. Review generated evaluations and rejection messages for privacy, employment-law, and bias considerations before using them in real hiring decisions.

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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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
Category
Not specified by scanner
Confidence
91% confidence
Finding

The README presents invocation examples like "Create a scorecard for Product Manager" and "Audit our interview process" as free-form triggers, but it does not define whether these are exact trigger phrases, examples only, or what contexts should or should not activate the skill. Because these are common business requests rather than narrowly scoped commands, they risk overlapping with ordinary conversation and causing unintended invocation.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The trigger phrase 'Evaluate this candidate' is broad and likely to overlap with normal conversation, making accidental invocation plausible. In an agent environment, ambiguous activation can cause the skill to run unexpectedly on sensitive hiring data, producing unintended processing or disclosure of candidate information.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The bias-mitigation section instructs interviewers not to treat culture fit as personal affinity and to rely on structured, evidence-based evaluation. The executive-hiring section then adds a board/exec team dinner for 'culture, not evaluation,' which reintroduces an unstructured social setting commonly associated with the kind of subjective culture-fit judgment the earlier guidance warns against.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The natural-language command set includes short, generic phrases that may be matched during ordinary discussion, increasing the risk of unintended skill activation. Because this skill handles hiring workflows and potentially sensitive employment data, accidental invocation could lead to unauthorized processing, logging, or generation of evaluative output without deliberate user intent.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.