lov-skill-optimizer
Pass
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/inspect_layout.pyusessubprocess.runto execute Git commands (rev-parse,status,branch) to gather repository information. These calls are implemented using list-based arguments to avoid shell injection vulnerabilities. - [INDIRECT_PROMPT_INJECTION]: The skill processes external files (SKILL.md, README.md, scripts) from other skills to perform audits and optimizations.
- Ingestion points:
scripts/lint_skill.pyandscripts/bump_version.pyread content from target skill directories. - Boundary markers: No explicit delimiters are used to wrap ingested content in the agent prompt, relying instead on the structured workflow steps.
- Capability inventory: The skill can write to the filesystem (
path.write_text,shutil.copy2) and execute Git commands (subprocess.run). - Sanitization: The tool uses regular expressions for analysis and modification rather than rigorous sanitization of input data.
- [DATA_EXFILTRATION]: The workflow includes
git push origin HEADto synchronize changes with a remote repository. This is an intended functional behavior for a version management tool and uses the user's existing Git configuration.
Audit Metadata