generate-changelog

Pass

Audited by Gen Agent Trust Hub on Sep 3, 2026

Risk Level: SAFEDYNAMIC_CONTEXT_INJECTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [DYNAMIC_CONTEXT_INJECTION]: The skill utilizes load-time shell execution via the !command syntax to populate the agent's context with current repository information. Commands include git status, git log, git describe, and git-cliff --unreleased. These are read-only operations relevant to the skill's primary function of version and changelog management.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests data from the repository's commit history, which serves as an untrusted input surface.
  • Ingestion points: Commit messages and history retrieved via git log and git-cliff in SKILL.md.
  • Boundary markers: No explicit delimiters or instructions are used to separate the ingested commit data from the agent's instructions.
  • Capability inventory: The skill possesses capabilities to write to the file system (CHANGELOG.md), create git commits, and create git tags.
  • Sanitization: No specific sanitization or filtering of commit message content is described prior to agent processing.
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute various CLI tools including git and git-cliff. These commands are used to analyze commits, modify local files, and manage repository tags, which aligns with the skill's intended use-case as a developer tool.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 3, 2026, 04:53 PM
Security Audit — agent-trust-hub — generate-changelog