Skills
skills/gihwan-dev/claude-code-gui/milestone-update/Gen Agent Trust Hub

milestone-update

Pass

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: LOWCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [Command Execution] (LOW): The skill uses git diff and git log to analyze session work. These are standard local operations and do not involve untrusted remote sources or shell piping.
  • [Indirect Prompt Injection] (LOW): The skill reads and processes the content of milestone.md and git logs to determine updates. 1. Ingestion points: Reads milestone.md and git history. 2. Boundary markers: Absent. 3. Capability inventory: File-write (to milestone.md) and git subprocess execution. 4. Sanitization: Absent. The skill assumes the content of project files is trustworthy, though the impact is restricted to local markdown updates.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 16, 2026, 12:33 PM