The Agent Skills Directory

[COMMAND_EXECUTION]: The skill extensively uses shell commands through Bash() calls and direct markdown instructions. This includes repository management with git, release orchestration with gh (GitHub CLI), and package management with npm and npx.
[EXTERNAL_DOWNLOADS]: The skill uses npx to execute the claude-flow package and its swarming components. This involves downloading and executing Node.js packages at runtime as part of its core functionality.
[PROMPT_INJECTION]: The skill exhibits vulnerability to Indirect Prompt Injection (Category 8):
Ingestion points: Reads untrusted commit messages and pull request metadata using gh api and gh pr list (found in SKILL.md).
Boundary markers: The instructions do not define clear delimiters or "ignore embedded instructions" warnings for the ingested commit and PR content.
Capability inventory: The skill possesses capabilities for file modification (Write in SKILL.md), shell execution (Bash in SKILL.md), and automated pull request creation (gh pr create).
Sanitization: There is no evidence of sanitization or filtering of the external content before it is processed by the AI for changelog generation or version suggestion.

github-release-management