The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from git commit logs to generate release summaries, which could contain malicious instructions designed to influence the agent's behavior during the release process.
Ingestion points: Step 1 in SKILL.md processes the output of git log --oneline -10 to infer changeset descriptions.
Boundary markers: No specific delimiters or warnings to ignore embedded instructions are provided when processing commit messages.
Capability inventory: The skill utilizes git, gh, npm, and npx to perform file writes, commits, pushes, and PR merges across all reference files and SKILL.md.
Sanitization: No sanitization or validation of the commit log content is performed before it is written to the .changeset/*.md file.
[COMMAND_EXECUTION]: The skill performs extensive shell command execution to manage the npm release pipeline.
Evidence: Executes git commit, git push, gh pr merge, and npm view as part of the core release automation logic defined in SKILL.md and the references/ directory.
[EXTERNAL_DOWNLOADS]: The skill references the installation and use of external tools from standard registries.
Evidence: Recommends running npm install -g npm@latest to ensure compatibility with trusted publishing workflows in references/version-pr-and-publish.md.

autoship