Skills
Navigate
skills/openharmonyinsight/openharmony-skills/oh-pr-workflow/Gen Agent Trust Hub

oh-pr-workflow

Fail

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: HIGHPRIVILEGE_ESCALATIONPERSISTENCECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [PRIVILEGE_ESCALATION]: The gitcode-mcp/install.sh script uses sudo to install the MCP server to system paths and modify file permissions.
  • [PERSISTENCE]: The gitcode-mcp/setup.sh script, which is automatically called by the skill during initialization if tools are missing, modifies agent configuration files (~/.claude.json and ~/.config/opencode/opencode.json) to register the server permanently.
  • [COMMAND_EXECUTION]: The skill performs local Go compilation and executes multiple bundled bash scripts to manage environment state and retrieve external data.
  • [EXTERNAL_DOWNLOADS]: The scripts/fetch_gate_defects.sh script makes network requests to dcp.openharmony.cn to retrieve defect reports.
  • [DYNAMIC_EXECUTION]: The scripts/fetch_gate_defects.sh script executes Python logic via a shell heredoc to handle complex API interactions.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted PR comments and CI reports, using their content to drive automated code modifications and repository writes.
  • Ingestion points: SKILL.md Mode 3 and Mode 5.
  • Boundary markers: None identified.
  • Capability inventory: git commit, git push --force, create_pr_comment, create_pull_request.
  • Sanitization: None.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Sep 16, 2026, 02:26 AM