The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands (pnpm and gt) to run type-checks, linters, and unit tests, and to manage pull request submission. These are standard operations for a development automation tool.
[SAFE]: The skill implements robust safety constraints and best practices to prevent unauthorized or high-risk changes. These include:
A strict limit of 100 lines of code for any single fix.
Explicit prohibitions against modifying sensitive system files related to authentication, database schemas (prisma/), and high-risk paths defined in the project's risk taxonomy.
A requirement to use draft pull requests (gt submit --draft), ensuring human review before any code is merged.
Mandatory verification steps requiring that a previously failing test must pass before submission.

bugfix