semantic-versioning

Pass

Audited by Gen Agent Trust Hub on Oct 13, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides Bash (bump-version.sh) and Python (bump_version.py) scripts that execute common development tools including git, npm, and jq. These scripts include case-matching and conditional logic to validate version increment types (major, minor, patch), which mitigates basic command injection risks.
  • [EXTERNAL_DOWNLOADS]: The documentation instructs the user to install well-known DevOps utilities such as semantic-release, standard-version, and git-cliff from established registries like NPM and Crates.io. These are standard industry tools for version management.
  • [CREDENTIALS_UNSAFE]: The provided GitHub Actions configuration demonstrates the correct use of encrypted secrets (GITHUB_TOKEN and NPM_TOKEN) for authentication, adhering to security best practices by avoiding hardcoded credentials.
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to analyze git commit history and project metadata (e.g., package.json, setup.py) to automate releases.
  • Ingestion points: Commit messages and package configuration files (SKILL.md, references/details.md).
  • Boundary markers: The documentation does not specify explicit boundary markers or delimiters for the automated tools to ignore adversarial instructions within commit messages.
  • Capability inventory: The skill uses tools that have the capability to write to the filesystem (npm version), commit changes, and create tags (git tag).
  • Sanitization: The provided scripts use strict string matching for version types, but the underlying third-party tools rely on parsing commit prefixes which could be manipulated by an attacker with commit access.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 13, 2026, 05:57 PM