firstbatchxyz

Inspect and analyze codebases using pygount for LOC counting, language breakdown, and code-vs-comment ratios. Use when asked to check lines of code, repo size, language composition, or codebase stats.

Set up GitHub authentication for the agent using git (universally available) or the gh CLI. Covers HTTPS tokens, SSH keys, credential helpers, and gh auth — with a detection flow to pick the right method automatically.

Production-grade PR review with execution-verified suggestions. Reads repository conventions, history, and security surfaces before reviewing. For every suggested fix, attempts to compile and test it in the sandbox — the comment includes proof. Modelled on GitHub Copilot's agentic architecture with one critical advantage: the sandbox is already running.

Create, manage, triage, and close GitHub issues. Search existing issues, add labels, assign people, and link to PRs. Works with gh CLI or falls back to git + GitHub REST API via curl.

Open and manage GitHub pull requests through Kai MCP tools — propose changes, monitor CI, iterate on failures, and merge. No git tokens are shared to the sandbox; every GitHub operation goes through the backend via the workspace's GitHub App installation.

Clone, create, fork, configure, and manage GitHub repositories. Manage remotes, secrets, releases, and workflows. Works with gh CLI or falls back to git + GitHub REST API via curl.

One-time repository onboarding for PR review — builds a RepoProfile (conventions, hot files, security surfaces, nit catalogue, test command) stored as workspace learnings. Re-runs automatically when the profile is stale (>30 days or >30% of files changed).

Manage AWS services for ML workflows including S3, ECR, SageMaker, and EC2

Tool surface for working in a adopt-bench synthetic GitHub organization. Use these when you need to read or modify the repository, list issues / PRs, open or update a PR, or comment on the team.

Turn a repo's house conventions into mechanical adopt-bench convention checks with REQUIRED fixture tests, then gate every generated task through lc-audit (the BenchGuard 2×2, cross-provider) before the owner approves it.

Stand up the run environment for a BYOR schedule — export it to a Harbor task, optionally wire an owner-supplied test_command for execution-verified grading, and prove it with an oracle smoke before any paid agent run.

Run a blind pairwise preference session over same-issue diffs from different agents/arms — "which PR would you merge?" — to produce an Elo/Bradley-Terry ranking and a human-vs-judge calibration number.

Build a adopt-bench learning-curve schedule from a repo the owner brings — interview them, mine issue↔merged-fix-PR pairs, synthesize leak-free prompts, calibrate difficulty with lc-ladder, and emit a schedule JSON ready for byor-checks.

Set up watchmen end-to-end for the current machine — install the CLI, run the init wizard, wire the plugin into Codex (and Claude Code), and verify the install with doctor. Use when someone wants to install, onboard, or configure watchmen with their coding agent.

Set up watchmen end-to-end for the current machine — install the CLI, run the init wizard, wire the plugin + statusline into Claude Code (and Codex), and verify the install with doctor. Use when someone wants to install, onboard, or configure watchmen with their coding agent.

Surface what watchmen detected in this workspace since the last curator run — new skills, AGENTS.md changes, suggested actions to take this session.

Surface what watchmen detected in this workspace since the last curator run — new skills, CLAUDE.md changes, suggested actions to take this session.