superbuild
Pass
Audited by Gen Agent Trust Hub on Mar 31, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill facilitates the execution of various development tools including linters, formatters, and test runners (e.g., npm test, pytest, go test, cargo test) through the provided scripts/validate-phase.sh utility. This script uses eval to run command strings constructed from hardcoded values based on detected project stacks.\n- [PROMPT_INJECTION]: Indirect Prompt Injection Vulnerability Surface. \n
- Ingestion points: The skill ingests untrusted implementation plans from the file system or user input (SKILL.md, Step 1: Ingest Plan). \n
- Boundary markers: The skill instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when processing plan content. \n
- Capability inventory: The agent has permissions to execute shell commands (via scripts/validate-phase.sh and tool calls) and modify local files (Step 5: Update Plan Document). \n
- Sanitization: No sanitization or validation of the plan document's instructions is performed before execution steps are carried out.
Audit Metadata