neumie
- 34 skills
- 0 followers
- 9 hours ago last updated
- ▌ Handoff · neumieUse when compacting the current conversation into a handoff document for another agent to pick up.
- ▌ Push · neumieUse when pushing commits to a remote. Checks tracking, shows what will be pushed, sets upstream if needed, handles diverged branches safely. Triggers: push, prepare for PR.
- ▌ Ship · neumieUse when shipping work end-to-end. Names branch, commits, pushes, creates PR — no confirmation. Triggers: ship, ship it, send it, go from uncommitted changes to open PR.
- ▌ Loop · neumie bundleUse when autonomously implementing a spec task-by-task. Each loop iteration picks one task, TDDs it, commits, exits. Progress persists in git. Triggers: loop, AFK mode, autonomous loop.
- ▌ Tdd · neumie bundleUse when writing code test-first via the red-green loop. Vertical-slice TDD at pre-agreed seams: one failing test, then minimal code to pass. Triggers: TDD, test-first, red-green.
- ▌ CI Fix · neumieUse when GitHub Actions CI is failing on the current branch. Fetches latest failed run, reads logs, finds root cause, fixes it. Triggers: fix CI, pipeline broken, build red.
- ▌ Commit · neumieUse when committing changes. Analyzes diff, writes a conventional commit message, commits immediately without confirmation. Triggers: commit, save my work, done with this change.
- ▌ Rebase · neumieUse when rebasing the current branch onto main or another base. Auto-detects base, handles conflicts, supports squashing. Triggers: rebase, sync with main, clean history before PR.
- ▌ Unslop · neumie bundleUse when prose sounds AI-generated or robotic and needs an audit, minimal cleanup, fact-preserving rewrite, or reconstruction in a taught human voice.
- ▌ Pr Watch · neumieUse when waiting for CI checks on a PR. Watches status, auto-fixes failures via ci-fix (max 2 retries), reports merge-ready. Triggers: watch PR, watch CI, wait for checks.
- ▌ Pr Create · neumieUse when creating a GitHub pull request. Pushes branch if needed, generates title + description from commits, creates PR via gh CLI. Triggers: create/open/submit PR.
- ▌ Implement · neumieUse when implementing one ready ticket or spec slice end to end. Verifies blockers, applies TDD where useful, reviews, updates queue state, and commits.
- ▌ Diagnose · neumie bundleUse when debugging hard bugs or perf regressions. Disciplined loop — reproduce, minimise, hypothesise, instrument, fix, regression-test. Triggers: debug this, broken, throwing, failing.
- ▌ Branch Name · neumieUse when naming or renaming the current git branch from its contents. Triggers: name this branch, rename branch, or working on default branch (main) with uncommitted work.
- ▌ To Tickets · neumieUse when breaking a plan, spec, or conversation into tracer-bullet tickets for GitHub or docs/plans, with blocking edges and agent/human readiness.
- ▌ Spec Create · neumieUse when turning a conversation or idea into docs/plans/<name>/spec.md. Synthesizes existing context into user stories, module design, testing decisions. Do NOT interview — just synthesize.
- ▌ Test Write · neumieUse when backfilling tests on existing code. Detects test framework, writes behavior-focused regression tests via public interfaces. Modes: targeted (files) or diff-based (changed).
- ▌ Code Review · neumie bundleUse when reviewing changes since a fixed point (branch, commit, tag, or PR base). Runs independent behavior, architecture, security, and verification/operations reviews.
- ▌ Commits Squash · neumieUse when squashing commits on the current branch into fewer logical commits. Triggers: squash, clean up commits, combine commits, tidy history before push or PR.
- ▌ Converge Loop · neumieUse when running converge-loop: repeat one custom command toward a mutable goal with overseer verdicts, steering, and stop/watch controls.
- ▌ Domain Model · neumie bundleUse when discussing codebase terminology, writing or editing CONTEXT.md glossary entries or ADRs, or when another skill needs domain-model maintenance.
- ▌ Explain · neumieUse when explaining code, concepts, systems, or supplied text in clear, context-appropriate technical language.
- ▌ Codebase Design · neumie bundleUse when designing module interfaces, seams, deep modules, testability, AI-navigable architecture, or when another skill needs design vocabulary.
- ▌ Prototype Build · neumie bundleUse when building throwaway logic or UI prototypes to answer design questions before production implementation; shareable HTML state demos or UI variants.
- ▌ Grill Me · neumieUse when stress-testing a plan, design, or architecture before writing a spec. Keeps resolved decisions in the conversation for spec-create.
- ▌ Grilling · neumieUse when stress-testing a plan, decision, or idea through dependency-aware rounds of grilling, especially for explicit "grill" requests.
- ▌ Codebase Improve · neumie bundleUse when finding architectural friction in a codebase. Surfaces shallow modules, proposes refactors for testability/AI-navigability, grills the design. Uses CONTEXT.md + ADRs.
- ▌ Wait What · neumieUse when the user says “wait what” or does not understand the immediately preceding substantive assistant response.
- ▌ Agents Md Map · neumieUse when mapping where CLAUDE.md or AGENTS.md files should exist. Scans for non-obvious complexity, evaluates existing files, identifies gaps. Run before agents-md-audit on each flagged file.
- ▌ Task Start · neumie bundleUse when starting a NEW task not yet explored. Assesses complexity, routes execution depth (trivial→solve, moderate→steps, complex→plan). Skip if task already discussed/planned.
- ▌ Agents Md Audit · neumie bundleUse when auditing a CLAUDE.md or AGENTS.md file. Scores six metrics (0-100), cites issues, suggests fixes. Codebase scan first, falls back to file-only when no codebase available.
- ▌ Branch Recap · neumieUse when recapping the current branch or picking up unfinished branch work: catch me up, what did I do, where was I.
- ▌ Grill With Docs · neumieUse when stress-testing a plan against the existing domain model + decisions. Challenges terminology, cross-references code, updates CONTEXT.md and ADRs inline as decisions crystallize.
- ▌ Complexity Assess · neumieUse when scoring how complex a task is before starting. Rates scope, clarity, risk, novelty into a tier (trivial/moderate/complex). Triggers: complexity check, how hard is this.