The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes multiple shell commands to interact with the local git repository and the GitHub CLI (gh). These commands are used to analyze commits, manage tags, and create releases.\n- [DATA_EXFILTRATION]: The skill transmits repository metadata and generated changelogs to external servers via git push origin and gh release create. These operations are transparent and consistent with the skill's purpose of publishing releases to remote platforms like GitHub.\n- [PROMPT_INJECTION]: The skill incorporates external, untrusted data from git commit messages into its output.\n
Ingestion points: Untrusted data enters the context through git log calls which capture commit subjects and bodies in SKILL.md.\n
Boundary markers: There are no explicit markers or instructions to the agent to ignore potentially malicious instructions embedded within commit messages.\n
Capability inventory: The skill has the capacity to execute shell scripts and perform network operations through git and GitHub tools.\n
Sanitization: There is no evidence of filtering or sanitizing the content of commit messages before they are added to the changelog and processed by the agent.

release