lodestar-audit
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill executes external commands including
gitandfallowusingchild_processmodules inscripts/audit-state.mjsandscripts/fallow-contract.mjs. These tools are used to analyze the repository state and source code as part of the audit process. - [DYNAMIC_EXECUTION]: The script
scripts/setup-modules.mjsperforms dynamic loading of modules from a sibling skill directory (lodestar-setup) usingimport(). This allows the skill to utilize shared logic from the lodestar tool suite. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted codebase content, creating an indirect prompt injection attack surface. The instructions in
SKILL.mdexplicitly mandate that repo content be treated as data/evidence only and that any steering instructions found in source files should be ignored. - [EXTERNAL_DOWNLOADS]: The skill includes logic to detect missing dependencies and recommends installation commands for
fallowusing standard package managers. It does not perform automated downloads or installations.
Audit Metadata