skills/yoshiko-pg/difit/release/Gen Agent Trust Hub

release

Pass

Audited by Gen Agent Trust Hub on Mar 30, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses command-line utilities such as git, npm, and gh to perform versioning tasks and manage remote repository state. These are standard tools used according to their intended purpose for release management.
  • [COMMAND_EXECUTION]: Executes a local script ./scripts/get-changes-since-tag.sh to retrieve the repository's commit history. This dependency assumes the script is present and safe within the user's environment.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests commit messages from the repository history to draft changelogs. An attacker could potentially include malicious instructions in a commit message. However, the requirement for explicit user confirmation in Japanese before any release actions occur serves as a critical safeguard.
  • Ingestion points: Commit data obtained from the repository history via local scripts.
  • Boundary markers: The skill does not define specific delimiters or isolation instructions for the processed commit content.
  • Capability inventory: File system access (updating CHANGELOG.md and package.json), versioning control (npm version), and network interaction (git push, gh release).
  • Sanitization: There is no automated sanitization or filtering of commit content before it is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 30, 2026, 03:26 AM