skills

Test-driven development via red-green-refactor with an auditable commit protocol. Use when implementing any feature, bugfix, refactor, or behavior change before writing implementation code, or when the user mentions TDD, red-green-refactor, or test-first development.

Terse chat that stays easy to parse. Drops filler, pleasantries and hedging while keeping full technical accuracy. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman.

Routes human attention to decisions that matter in agent-generated code. Active during planning, implementing, fixing. Defines when and how to place DECISION markers in code comments. Also applies when reviewing a diff/PR on request.

Build and sharpen a project's domain model. Use when the user wants to pin down domain terminology or a ubiquitous language, record an architectural decision, or when another skill needs to maintain the domain model.

Grill the user about a plan, decision, or idea via scored multiple-choice questions. Use when the user wants to stress-test their thinking or uses any 'grill' trigger phrase.

Prepare a session for compaction: write the handoff, mark it fresh, propose /compact focus lines. Use before a manual /compact, when a reminder or hook warns that compaction is near, or when asked to wrap up or hand off session context.

Run a fresh-session probe against this environment from a checked-in probe reference. Use when the user asks to run a probe, reinset run, or live-fire dojo test of session machinery, or names a probe dir.

Execute a checked-in probe fixture in this session. Use when a message asks to use take-probe with a reference, or hands this session a probe dir to execute.

Records a session's open work as append-only events in the session-memory store, and renders the watchable view. Use when a session's progress must be visible outside it — an orchestrator tracking parallel threads, or a worker or review run reporting where it got to.

Turn the current conversation into a spec published to the project tracker — no interview, just synthesis of what was already discussed.

Break a plan, spec, or the current conversation into tracer-bullet tickets with blocking edges, published to the project tracker. Use when splitting approved work into issues.

Format, numbering, and when-to-write criteria for Architecture Decision Records in docs/adr/. Use when recording an architectural decision, when another skill flags a decision as ADR-worthy, or when the user asks to write or review an ADR.

Prose rules per surface, and the loop that writes, rewrites and reviews prose against them. Use when writing or rewriting docs, code comments, commit messages, tickets, pull request bodies, review comments, skill files or CLAUDE.md, or when reviewing any of them for prose.

Write and edit agent skills that take the same process every run. Use when authoring a skill, editing or pruning an existing one, deriving a skill from an upstream source, or reviewing a change to a skill file.