any-percent
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell commands for project management and resource cleanup.
- Evidence: Specifically invokes
git worktree removeandgit branch -Dto clean up variant implementations and branches. - [DYNAMIC_EXECUTION]: The workflow relies on generating and executing code based on LLM output.
- Evidence: Instructs the agent to generate first-pass code variants and run automated tests to verify them.
- [INDIRECT_PROMPT_INJECTION]: The skill processes code from an external source, which represents an attack surface for instructions embedded in data.
- Ingestion points: Code generated by the
mcp__speed-run__generate_and_write_filestool. - Boundary markers (absent): No specific delimiters or warnings are provided to the agent to ignore instructions embedded within the generated code.
- Capability inventory: The agent can execute shell commands for Git operations and run automated test suites.
- Sanitization (absent): There is no verification or sanitization process for the code generated by the hosted LLM before it is written to the filesystem or executed.
Audit Metadata