trellis-spec-bootstarp

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the installation of abcoder from CloudWeGo's GitHub repository (github.com/cloudwego/abcoder). CloudWeGo is a well-known open-source organization.
  • [EXTERNAL_DOWNLOADS]: The skill uses npx gitnexus to analyze codebases and provide an MCP server, which downloads and executes the package from the npm registry.
  • [COMMAND_EXECUTION]: The skill involves executing several shell commands for environment setup and verification, such as go install, npx gitnexus analyze, abcoder parse, and grep searches for placeholders.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted codebase data to generate documentation, presenting an indirect prompt injection surface.
  • Ingestion points: Direct source reads, GitNexus graph queries, and ABCoder AST parsing results in references/repository-analysis.md and references/mcp-setup.md.
  • Boundary markers: The instructions do not specify the use of delimiters or specific warnings to ignore embedded instructions within the analyzed codebase.
  • Capability inventory: The agent can write files to the .trellis/spec/ directory and execute shell commands via the command line.
  • Sanitization: There is no evidence of sanitization or filtering applied to the code samples or metadata extracted from the repository before they are incorporated into the generated specifications.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 08:21 AM