Back to skill

Security audit

声音制作规范,Jiuge_Flow_Perfect_V1.skill

Security checks for vulnerabilities and agentic risk

Overview

This audio-processing skill is purpose-aligned overall, but its script can operate outside the intended input and output folders when given a path-like filename.

Review this skill before installing. Use it only with simple filenames already placed in Jiuge_Input, avoid sensitive recordings on shared systems, and prefer a fixed version that rejects path separators and traversal before creating output directories or running cleanup.

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 (1)

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/jiuge_flow.sh:17
Finding

Unvalidated Input Filename Enables Path Traversal Outside Approved Directories

Content
View full analysis
/dev/null || true log_success "Cleanup completed" } ``` ### Technical Analysis `INPUT_FILE` is supplied directly by the caller and is not restricted to a simple filename. Although its shell expansions are generally quoted, quoting only prevents shell word splitting and ordinary shell metacharacter injection. It does not prevent filesystem traversal using absolute paths or components such as `../`. The input is appended to `INPUT_DIR` for the existence check and is also used to derive `PROJECT_NAME`. As a result, traversal components can affect both the source file and the destination project path: ```bash "${INPUT_DIR}/${INPUT_FILE}" "${OUTPUT ...[truncated 2368 chars]
Remediation
View remediation
Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill states that processed files are automatically archived, but the description does not clearly warn users about this behavior at the point where they are likely to decide whether to use the skill. Automatic archiving can expose sensitive audio content, metadata, or project names to unintended retention or discovery, especially on shared systems or when users assume outputs remain only in the working directory.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
78% confidence
Finding

The script's natural-language interface and status text are entirely in Chinese, starting from the header description and continuing through the user-facing workflow messaging. Under the stated policy, forcing a specific language without opt-in can be a locale-policy issue when no alternative or language selection is documented.

Content

No source excerpt is available for this finding.