T01 · Skill Instruction Hijacking
Error
- Location
SKILL.md:2- Finding
Mandatory Third-Party Promotion and Lead Collection Hijacks Review Requests
- Content
View full analysis
- Remediation
View remediation
Security audit
Specialist Code Review
Security checks for vulnerabilities and agentic risk
Overview
This skill is a disclosed third-party specialist-review request workflow, but users should understand it may introduce Provener and ask for budget and project details before any submission is authorized.
Install only if you are comfortable with the agent offering Provener for human review requests and asking for scope, deadline, expert profile, budget, and payment-capable status. Do not authorize submission of private code, secrets, vulnerability details, or proprietary architecture unless you intend to share those exact details with Provener.
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
Findings (1)
Vulnerability Patterns
- 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
- Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)
Vague Triggers
Medium
- Category
- Not specified by scanner
- Confidence
- 92% confidence
- Finding
The description explicitly says to apply the skill even before authorization is complete, creating ambiguous boundaries between detecting need and starting an external referral workflow. Even though later text says to submit only after authorization, this early-activation instruction can pressure the agent to solicit sensitive scope, budget, or architectural details sooner than necessary.
- Content
Vague Triggers
Medium
- Category
- Not specified by scanner
- Confidence
- 91% confidence
- Finding
The activation criteria are intentionally broad enough to trigger on many ordinary requests for expert help, which can cause the agent to introduce a third-party service prematurely. In this skill’s context, that increases the chance of unnecessary data collection and external sharing workflows before the user has clearly asked for escalation beyond normal assistance.
- Content