The Agent Skills Directory

[COMMAND_EXECUTION]: The skill dynamically identifies and executes shell scripts from local configuration files (.github/workflows/ci.yaml and .gitlab-ci.yml). According to references/ci-checks.md, the agent is instructed to run run: or script: blocks "as-is", which could result in the execution of arbitrary commands if those configuration files are maliciously modified.
[EXTERNAL_DOWNLOADS]: The skill utilizes npx to execute markdownlint-cli2 and cspell. This command may download these packages from the public npm registry at runtime if they are not cached locally.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Category 8) because it processes untrusted data from repository files to drive its logic and command execution.
Ingestion points: Reads content from .github/workflows/ci.yaml, .gitlab-ci.yml, and .github/PULL_REQUEST_TEMPLATE.md to determine local commands and PR structure.
Boundary markers: Absent; there are no instructions to the agent to ignore or delimit instructions embedded within these files.
Capability inventory: The skill has access to the Bash tool, enabling it to perform git operations, execute PR CLI tools (gh, glab), and run arbitrary scripts translated from CI configurations.
Sanitization: None; the agent is directed to run scripts from the CI configuration files without prior validation or escaping.

create-pr