init-or-refactor-agents-md
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill reads and processes the contents of repo-local instruction files such as
AGENTS.md,CLAUDE.md,CODEX.md, and.cursorrules. These files could potentially contain malicious instructions embedded by external contributors (e.g., via a pull request) designed to influence the agent's behavior during the refactoring process. - Ingestion points: Workflow steps 1 and 2 in
SKILL.mdspecify reading multiple local instruction files. - Boundary markers: There are no explicit instructions or delimiters used to separate these inputs from the agent's system instructions or to treat them as untrusted data.
- Capability inventory: The skill has the capability to write and modify files (
AGENTS.mdandCLAUDE.md) based on the processed content. - Sanitization: No sanitization or validation of the input content is specified in the workflow.
Audit Metadata