ziniao-shared

Pass

Audited by Gen Agent Trust Hub on Aug 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
  • [PERSISTENCE_MECHANISMS]: The skill specifies instructions to modify the user's shell profile (~/.bashrc) by appending export MSYS_NO_PATHCONV=1. This is intended to resolve specific path conversion issues encountered in Windows Git Bash environments.
  • [DATA_EXPOSURE_AND_EXFILTRATION]: The skill documents that sensitive credentials such as apiKey and companyId are stored in ~/.ziniao-cli/config.json and the system Keychain. It includes a security rule instructing the agent to avoid outputting full API keys to the terminal to prevent credential exposure.
  • [UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The skill references updating the CLI tool via npm update -g @ziniao-open/cli. The package is a vendor-owned resource from the author ziniao-open.
  • [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to parse tool output to drive subsequent actions. This creates a vulnerability surface if the processed output contains malicious instructions.
  • Ingestion points: Tool output streams (stdout/stderr) from ziniao-cli commands.
  • Boundary markers: Absent. No delimiters are used to wrap or identify untrusted content from the tool.
  • Capability inventory: Execution of shell commands, network operations via the CLI, and filesystem access.
  • Sanitization: Absent. No filtering or validation of tool-generated strings is performed before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 17, 2026, 03:44 PM
Security Audit — agent-trust-hub — ziniao-shared