The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes multiple shell commands using git and gh to manage the development workflow, including branch creation, pushing commits, and interacting with GitHub pull requests.
[EXTERNAL_DOWNLOADS]: During worktree setup and lockfile reconciliation, the skill may invoke pnpm install to ensure that necessary dependencies are available in the local environment.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data such as git commit messages and branch names to automatically generate PR titles and changelog content.
Ingestion points: Commit messages are retrieved via git log and branch names via git branch, which are then parsed by the derive-bump.mjs script for processing by the agent.
Boundary markers: The skill does not explicitly use delimiters or specialized instructions to prevent the agent from potentially following instructions that might be embedded within commit messages or PR metadata.
Capability inventory: The agent has access to file system tools (Write, Read, Edit) and shell execution capabilities for developer tools (Bash(git:*), Bash(gh:*), Bash(pnpm:*)).
Sanitization: While the derive-bump.mjs helper script parses the commit structure according to Conventional Commits standards, it does not sanitize the text content for secondary instructions before the agent processes it for PR generation.

send-it