context-labs

Playbook for building a NET-NEW feature in whip. Use when the user wants to add a feature, tool, slash command, integration, or UX behavior ('build…', 'add…', 'implement…'), even without the word 'feature'. Mines docs/roadmap.md + docs/learnings/. NOT for…

The lazy senior dev: write the least code that works — the ponytail ladder (does it need to exist? already here? stdlib? native? existing dep? one line?), root cause over symptom, deletion over addition, `ponytail:` comments for shortcuts. whip's default…

Golang benchmarking, profiling, and performance measurement. Use when writing or comparing Go benchmarks, profiling with pprof, interpreting CPU/memory/trace profiles, benchstat analysis, CI benchmark regression detection, or production performance issues.…

Golang CLI development — command structure, flags, config layering, exit codes, I/O, signals, shell completion, validation, testing. Use when building or reviewing a Go CLI or code using cobra, viper, urfave/cli. Cobra → `golang-spf13-cobra`; viper →…

Golang code style — line length and breaking, variable declarations, control flow clarity, when comments help vs hurt. Use when writing or reviewing Go code or setting coding standards. Not for naming (→ `golang-naming`), lint (→ `golang-lint`), docs (→…

Golang concurrency patterns. Use when writing or reviewing concurrent Go code — goroutines, channels, select, locks, sync primitives, errgroup, singleflight, worker pools, fan-out/fan-in — or for goroutine leaks, race conditions, channel ownership, or…

Idiomatic context.Context usage in Golang — cancellation, timeouts/deadlines, request-scoped values, WithoutCancel for background work. Use when designing context propagation across layers, debugging leaked contexts, or choosing Background/TODO/WithoutCancel.…

CI/CD with GitHub Actions for Golang — testing, linting, SAST, security scanning, coverage, Dependabot, Renovate, GoReleaser, release pipelines. Use when setting up Go project CI, GitHub Actions workflows, linters/scanners, dependency update automation, or…

Adversarial multi-model code review that finds real defects, proves them, and fixes them. Every available model family (grok, codex, claude — including the one hosting this session) reviews the diff independently and blind; every finding is then…

The disciplined playbook for fixing a bug in the clocked-in repo — reproduce and root-cause BEFORE changing code, prove it with a failing test, fix at the right layer, verify the gate, then clean up. Use whenever something is broken or wrong ("X isn't…

Shape output for a reader with ADHD: lead with the next action, number multi-step work, restate state across turns, suppress tangents, give specific time estimates, make wins visible. Invoke with /i-have-adhd; stays on until "stop adhd mode".

Vendor-neutral routing doctrine for delegating implementation work across the Claude Code, Grok, and Codex CLIs. Whichever model hosts the session acts as architect and routes typing to the other two families. Use when delegating implementation, choosing an…

The end-to-end playbook for building a NET-NEW feature in the inference monorepo — the umbrella that turns "I want to build X" into shipped, observable, well-tested, service-pattern-compliant code. Use this whenever the user wants to add a feature, endpoint,…

How to run and debug tests in the clocked-in repo — a small Bun CLI. Covers `bun test`, the `task check` gate (tsgo + oxlint + oxfmt + tests), running a single file or test, how DB/HOME isolation works, the committed-bundle rebuild gotcha, and end-to-end…

Use HALO Engine as a diagnostic aid to iteratively improve an agent harness. Engine surfaces patterns from traces; Claude maps findings to code, makes the smallest sensible change, re-runs, and measures.

Use the lsptkns CLI for semantic code analysis via Language Server Protocol. Use this skill whenever you need to find symbol definitions, trace references across a codebase, list all symbols in a project, inspect class/function internals, build call graphs,…