agents-md-improver

Pass

Audited by Gen Agent Trust Hub on Mar 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill manages local repository files including AGENTS.md, CLAUDE.md, .claude/CLAUDE.md, and GEMINI.md through read, write, delete, and symlink operations. These actions are restricted to the local repository and require explicit user confirmation before execution.
  • [PROMPT_INJECTION]: The skill is designed to translate user input into persistent repository instructions, creating an indirect prompt injection surface. * Ingestion points: User-provided corrections and instruction update requests in SKILL.md. * Boundary markers: Not explicitly defined for the generated patch content; however, the skill mandates a 'propose-first' confirmation step. * Capability inventory: File-system access (read, write, delete, symlink) within the repository as defined in SKILL.md. * Sanitization: No explicit sanitization or validation of user input is described before it is incorporated into the instruction files.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 15, 2026, 09:10 PM