code-refine
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill suggests installing a companion skill,
code-refinements, from the author's GitHub repository using the commandnpx skills add https://github.com/alexandru/skills --skill code-refinements. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted code and project instructions which may contain malicious patterns intended to influence the agent's behavior during the refinement process.
- Ingestion points: Workspace source code, branch diffs, and project-specific instructions within the
SKILL.mdfile and project files. - Boundary markers: None identified.
- Capability inventory: File system modifications, test execution, and companion skill installation via shell commands.
- Sanitization: No explicit sanitization or validation of the ingested code or instructions is described.
Audit Metadata