NYC

release-notes-composer

Warn

Audited by Gen Agent Trust Hub on Feb 15, 2026

Risk Level: MEDIUMPROMPT_INJECTIONNO_CODE
Full Analysis
  • [Indirect Prompt Injection] (MEDIUM): The skill is susceptible to indirect prompt injection via untrusted git commit history. * Ingestion points: Git commit messages and repository history are processed as inputs. * Boundary markers: Absent; there are no instructions to delimit or ignore embedded commands within the commit data. * Capability inventory: The skill reads repository history and suggests writing results to the local filesystem (CHANGELOG.md). * Sanitization: None identified. A malicious actor could craft a commit message containing instructions that the LLM might follow, such as including phishing links or altering the summary logic.
  • [No Executable Code] (SAFE): No executable scripts or dependency manifests (e.g., package.json, requirements.txt) were provided, which eliminates immediate risks of Remote Code Execution or local file exfiltration through code execution.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 15, 2026, 10:05 PM