The Agent Skills Directory

[COMMAND_EXECUTION]: Executes a local helper script (scripts/log-bug.js) to process JSON metadata into CSV format for audit logging. This is a functional requirement of the skill.
[COMMAND_EXECUTION]: Invokes standard development CLI tools such as git, gh (GitHub CLI), npm, and tsc to perform version control operations and run test suites during the verification phase.
[EXTERNAL_DOWNLOADS]: Interacts with GitHub's official services via the gh command-line tool to manage bug reports and sync issues, which represents expected developer workflow behavior.
[PROMPT_INJECTION]: Provides a surface for indirect prompt injection as the skill is designed to process external data like bug reports and terminal error traces.
Ingestion points: Processes user-provided bug descriptions, terminal output, and temporary data files (e.g., bug.json).
Boundary markers: Not explicitly defined in the logic; the skill relies on the agent's ability to distinguish between data and instructions during triage.
Capability inventory: Accesses the filesystem for CSV logging and uses gh CLI for repository interaction.
Sanitization: The log-bug.js utility performs standard CSV escaping by sanitizing newlines and escaping double quotes in input strings.

fix-report