T09 · Insecure Skill Coding Practices
Warning
- Location
backup-restore.py:16- Finding
Unencrypted Storage of Potentially Sensitive Backup Data
- Content
View full analysis
- Remediation
View remediation
Security audit
Backup Restore
Security checks for vulnerabilities and agentic risk
Overview
This backup skill is review-worthy because it creates local backups but overstates restore, scheduling, and coverage, including a restore command that reports success without restoring anything.
Treat this skill as unfinished. Do not rely on it for disaster recovery or migration unless you first fix and test restore behavior, source-path selection, backup completeness, encryption or storage protections, and command-line parsing. If installed as-is, use it only for non-critical manual experiments and protect any ZIP files it creates as sensitive data.
Vulnerability Patterns
- Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
- 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
Findings (4)
T09 · Insecure Skill Coding Practices
Warning
- Location
backup-restore.py:70- Finding
Restore Operation Reports Success Without Restoring Data
- Content
View full analysis
- Remediation
View remediation
T09 · Insecure Skill Coding Practices
Warning
- Location
backup-restore.py:81- Finding
Backup Verification Does Not Establish Authenticity or Completeness
- Content
View full analysis
- Remediation
View remediation
T09 · Insecure Skill Coding Practices
Note
- Location
backup-restore.py:16- Finding
Hard-Coded User Path and Incomplete Backup Scope Can Cause Silent Data Omission
- Content
View full analysis
- Remediation
View remediation
Vulnerability Patterns
- 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
- Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)
Tp4
High
- Category
- MCP Tool Poisoning
- Confidence
- 95% confidence
- Finding
The skill advertises backup, automatic scheduling, and one-click restore coverage that is broader than what the available description indicates is actually implemented. In a backup/restore context, this is dangerous because operators may rely on incomplete backups or assume restoration capability exists during an incident, leading to data loss or failed recovery when it matters most.
- Content
Intent-Code Divergence
High
- Category
- Not specified by scanner
- Confidence
- 98% confidence
- Finding
The restore function explicitly claims to perform a restore, but it only checks whether the backup file exists and then prints success without extracting or reinstating any data. In a recovery scenario, this deception is especially dangerous because it can lead users to believe restoration succeeded while systems remain unrecovered, potentially worsening downtime and data loss.
- Content
Missing User Warnings
Medium
- Category
- Not specified by scanner
- Confidence
- 93% confidence
- Finding
Documenting restore commands without prominent warnings about overwrite/destructive behavior can cause users to replace existing configuration, workspace, or agent state unintentionally. Because this skill targets system backup and restore, the context increases risk: users are likely to run these commands during emergencies and may not realize they need confirmation, dry-run support, or isolation before restoring.
- Content
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 90% confidence
- Finding
The file contains user-facing natural-language strings such as the title, status messages, and usage instructions exclusively in Chinese. Under the policy rule, forcing a specific language without user opt-in is a natural-language policy violation unless the locale restriction is explicitly justified.
- Content
Description-Behavior Mismatch
Medium
- Category
- Not specified by scanner
- Confidence
- 94% confidence
- Finding
The skill description promises backup coverage for skills and one-click restore, but the implementation only archives a few config/workspace paths and does not include skills or any working restore logic. This mismatch can create a false sense of recoverability, causing operators to rely on backups that will not restore the expected assets after data loss or corruption.
- Content
Natural-Language Policy Violations
Low
- Category
- Not specified by scanner
- Confidence
- 88% confidence
- Finding
The natural-language description forces a specific language/locale for core skill instructions and labels, and the file does not indicate that Chinese is optional or required for a region-specific purpose. Per the policy, language constraints should offer user choice or be explicitly justified.
- Content
Static analysis
No suspicious patterns detected.