Changelog Management

Pass

Audited by Gen Agent Trust Hub on Feb 27, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local Git and PowerShell commands to manage repository metadata.
  • Evidence: Uses git tag -l, git log, and PowerShell's Select-String to identify version boundaries and extract commit history.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted data from Git commit logs.
  • Ingestion points: Commits retrieved via git log (Workflow Step 2) are processed to generate changelog entries.
  • Boundary markers: Absent; there are no delimiters or instructions to the agent to ignore instructions embedded within commit messages.
  • Capability inventory: The skill has permissions to read and write local files (CHANGELOG.md, Directory.Build.props) and execute Git CLI commands.
  • Sanitization: Absent; commit messages are parsed and consolidated directly into the final output without filtering for malicious instructions.
  • [PROMPT_INJECTION]: The skill uses instructional markers to enforce specific task logic.
  • Evidence: Uses "CRITICAL" labels to ensure the agent prioritizes git tag checking and commit consolidation over other formatting rules.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 27, 2026, 01:09 AM