The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs automated tasks by executing shell commands including git, npm, and mvn. These commands are used to modify project manifests, commit changes, create tags, and push code to remote repositories.
[INDIRECT_PROMPT_INJECTION]: The skill analyzes external, untrusted content such as git commit messages and the CHANGELOG.md file to determine the next version level (e.g., detecting 'feat' or 'breaking change' in history).
Ingestion points: The agent is instructed to read the root CHANGELOG.md and user-facing commits since the last tag (using git log or git describe) in SKILL.md Section 2.
Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions when processing the content of commit messages or changelog entries, which could contain malicious payloads designed to influence agent behavior.
Capability inventory: The skill allows the agent to modify files (package.json, pom.xml, gradle.properties, pyproject.toml, etc.), commit changes, and perform network operations via git push to origin.
Sanitization: No explicit sanitization or validation logic is defined for the strings parsed from the git history or changelog before they are processed by the agent.

version-release