agents-md
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to inspect repository configuration files, manifests, and documentation (such as
package.json,README.md,CONTRIBUTING.md, and task runners) to synthesize project instructions intoAGENTS.md. This processing of repository-supplied files introduces an indirect prompt injection surface if an untrusted or compromised repository file contains adversarial instructions. - Ingestion points: Local repository metadata, manifests, and documentation files parsed during the initial inspection workflow (
SKILL.md). - Boundary markers: Absent. There are no explicit boundaries or guidelines directing the agent to disregard natural language instructions embedded within the analyzed project files.
- Capability inventory: Limited to file creation and modification of
AGENTS.mdand symlinkingCLAUDE.md. The skill includes no capabilities for active code execution, process spawning, or external network operations. - Sanitization: Absent. Content from existing files is summarized and structured without active text sanitization or validation controls.
Audit Metadata