The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill executes bunx oxfmt, which downloads and runs the oxfmt package from the npm registry at runtime. While oxfmt is a recognized tool in the JavaScript ecosystem, it is an external dependency managed at execution time.
[COMMAND_EXECUTION]: The skill performs several command-line operations, including bun run build, bun run stylecheck, git push, and various gh (GitHub CLI) commands. These actions are consistent with the skill's stated purpose of preparing and submitting a pull request.
[PROMPT_INJECTION]: Indirect Prompt Injection Surface (Category 8):
Ingestion points: The skill reads the local feature work and branch status (via gh pr view/status) to determine if a PR is needed and to generate the PR title and body.
Boundary markers: There are no explicit instructions or delimiters used to ensure the agent ignores potentially malicious instructions embedded in the local source code or issue descriptions being processed.
Capability inventory: The skill has the ability to execute shell commands, perform network operations (git push, gh pr create), and write to the filesystem (/tmp/).
Sanitization: The skill effectively mitigates shell command injection by writing the PR body to a temporary file (/tmp/remotion-pr-body.md) and using the --body-file flag instead of passing user-controlled content as an inline shell argument.

pr