devotts

@devotts source repo

32 skills
0 followers
7 hours ago last updated

▌
Review It 2 · devotts
The QA front door of the DevOtts lifecycle family — plan-it plans, build-it builds, /review-it verifies. PRIMARY mission: run the unit tests, e2e tests and test-cases generated at plan phase to prove the build obeys the plan — the independent verification leg of the plan→build→review triangle. Also verifies third-party side-effects (the Airtable class), staging/prod deploys (deployed-code ladder + [REAL] re-runs), and runs a severity-tiered PR review. Routes execution to full-qa, iterate, chrome-cdp-control, make-eval and parallel-lifecycle — never re-implements them — and enforces an 11-rule gate catalog that makes false-VERIFIED claims un-shippable. Invoked with no Test Contract it never refuses and never self-grades — it runs the no-contract ladder and tags every verdict AUTHORED or DERIVED. Use when the user says "/review-it", "review it", "verify the build", "run the test contract", "QA this feature", "verify this deploy", "review this PR", or when build-it reaches its QA phase.

0
installs
▌
Conclude It 2 · devotts bundle
Conclude It

0
installs
▌
Prompt It 2 · devotts bundle
Harness-aware intent compiler: turns a rough ask into an optimized, context-grounded prompt — routed to the right *-it skill with exactly the slots that target parses, nothing it already owns. Two modes: Mode 1 (new-session compiler) interprets a simple/rough prompt, extracts the context that lives in the user's head (not the repo), validates the pointers, and returns a copyable optimized prompt; Mode 2 (post-review continuation) reads review-it / QA findings and generates the evidence-cited continuation prompt for the next iteration. Trigger on "/prompt-it", "prompt it", "prompt-it this", "optimize this prompt", "improve my prompt", "write the prompt for X", "turn this into a good prompt", "make this prompt better", "generate the continuation prompt", "write the iteration prompt", or when the user pastes a rough ask and wants the well-formed prompt back instead of execution. NOT for post-planning handoffs of a finished PRD/spec — that is /next-session-prompt; prompt-it compiles NEW intent (Mode 1) or post-RE

0
installs
▌
Skill It 2 · devotts
Package a Claude/agent skill (or a whole plugin) into a public, installable, marketing-ready repo — the exact recipe behind the DevOtts *-it family (plan-it, build-it, review-it, prompt-it, conclude-it). Name workshop → genericization (two-pass sweep proving zero personal references) → progressive-disclosure structure → standard scaffold (root SKILL.md crawler manifest, plugin + marketplace manifests, pain-first marketing README + hero SVG + per-agent install matrix + compatibility table, CHANGELOG, QA, MIT LICENSE) → mechanical validation → ship (gh repo create → main → topics → tagged release) → installability probe. Trigger on "/skill-it", "skill it", "skill-devotts-builder", "package this skill", "ship this skill as a plugin", "create the plugin repo for X", "publish my skill like build-it", "turn this folder into an installable skill repo".

0
installs
▌
Build It 2 · devotts
Single-command autonomous delivery orchestrator. Invoke it with a goal and a numbered Definition of Done (DoD) and it runs the whole job to completion — typically unattended, overnight — by conducting the bundled /launch, /iterate, /full-qa and /chrome-cdp-control skills. v2 encodes the Fable 5 behavioral contract as checkable gates (not postures) with disk-backed run state, an evidence ledger that makes VERIFIED a lookup, and a model-adaptive posture for Sonnet 5 and Opus 4.8. Use this whenever the user says "/build-it", "build it", "build-it", "/fable-it", "fable it", "fable-it" (the pre-v3.1.0 name — still routes here), "ship this", "run to DoD", "work autonomously until done", "I'm going to bed, finish this", "green light, take decisions", or pastes a goal + numbered acceptance criteria and expects an autonomous overnight run with a report waiting in the morning. Also use when the request describes an agile, cycle-based build (epics → tests → code → test → fix → loop) split across Claude teams or subagent

0
installs
▌
Notebooklm · devotts
Complete API for Google NotebookLM - full programmatic access including features not in the web UI. Create notebooks, add sources, generate all artifact types, download in multiple formats. Activates on explicit /notebooklm or intent like "create a podcast about X"

0
installs
▌
Prompt It · devotts bundle
Harness-aware intent compiler: turns a rough ask into an optimized, context-grounded prompt — routed to the right *-it skill with exactly the slots that target parses, nothing it already owns. Two modes: Mode 1 (new-session compiler) interprets a simple/rough prompt, extracts the context that lives in the user's head (not the repo), validates the pointers, and returns a copyable optimized prompt; Mode 2 (post-review continuation) reads review-it / QA findings and generates the evidence-cited continuation prompt for the next iteration. Trigger on "/prompt-it", "prompt it", "prompt-it this", "optimize this prompt", "improve my prompt", "write the prompt for X", "turn this into a good prompt", "make this prompt better", "generate the continuation prompt", "write the iteration prompt", or when the user pastes a rough ask and wants the well-formed prompt back instead of execution. NOT for post-planning handoffs of a finished PRD/spec — that is /next-session-prompt; prompt-it compiles NEW intent (Mode 1) or post-RE

0
installs
▌
Skill It · devotts bundle
Package a Claude/agent skill (or a whole plugin) into a public, installable, marketing-ready repo — the exact recipe behind the DevOtts *-it family (plan-it, build-it, review-it, prompt-it, conclude-it). Takes a local skill folder (or an idea for one), genericizes it (two-pass sweep proving zero personal/project references), applies progressive disclosure (thin SKILL.md + references/ legs), scaffolds the standard repo (root SKILL.md crawler manifest, Claude Code plugin + marketplace manifests, marketing README with a pain-first headline + hero SVG + per-agent install matrix + 18-row compatibility table, CHANGELOG, QA walkthroughs, MIT LICENSE), validates everything mechanically (JSON/YAML/SVG parse, version consistency, script smoke tests), then ships: gh repo create → main → discovery topics → tagged release with zip — and verifies installability with the skills CLI before claiming anything. Trigger on "/skill-it", "skill it", "package this skill", "ship this skill as a plugin", "create the plugin repo for m

0
installs
▌
Conclude It · devotts bundle
The single front door for every Claude Code session close — and the ship stage of the *-it lifecycle (plan-it plans it, build-it builds it, review-it reviews it, conclude-it concludes it). A mechanical Step-0 triage detects ship-shaped sessions (deploys run, feature work merged, DoD met without prod verification) and runs the ship pipeline first: local test gate → merge+deploy → prod gate → archive plan docs → backlog → PM board. Then the session-close core every session gets: knowledge debrief with a supersession sweep (old docs get updated or deprecated, never left "current"), session-history ledger card, mid-flight + controlled-sessions verification, idle-subagent reaping, own-footprint sweep including secrets at rest, a local CONCLUDE_REPORT.html double-check report, and an honest safe-to-exit verdict. Fail-closed gates, no fake greens. First run in a repo: a five-question setup interview (docs home, second brain + scope, deploy skill, debrief skill) saved INSIDE the repo — once per repo, global only on e

0
installs
▌
Session Debrief · devotts
End-of-session knowledge capture for any project — updates the project's documentation folder, repo skills, and (optionally) the user's knowledge vault with the patterns, gotchas, and fixes discovered during the session. Includes the anti-rot promotion rule (durable pages updated + verified-date bumped, never just session notes) and the supersession sweep (pages describing the OLD behaviour get updated or deprecated — never left "current"). Invoked by /conclude-it Step 2; also usable standalone at the end of any working session ("debrief the session", "capture what we learned").

0
installs
▌
Plan It · devotts bundle
Turn a fuzzy idea, brain-dump, or transcription into a COMPLETE spec set + agile delivery package for /build-it. /plan-it plans it, build-it builds it: pre-grounds the codebase, fans out parallel Claude teams, authors design docs in order, runs one up-front anamnesis questionnaire (access, fences, naming, topology, live-probes, known decisions), pauses at ONE batched human-decision gate, then freezes a shared CONTRACT so squads write PRDs + epics — each ending in a BINDING Test Contract the build must pass 100% before "done". Picks a topology (solo · orchestrator+squads · headless) and renders an HTML twin of each doc. Use when the user says "/plan-it", "plan this", "spec this out", "create the PRDs/epics", "scope this project/feature", or pastes a vision expecting a buildable plan. Built for humans and conductor agents. Inverse of /build-it; predecessor to /next-session-prompt. Deterministic core: an explicit statechart (machine.json), a resumable state file, gate-check.mjs exit codes gate advancement.

0
installs
▌
Sample Skill · devotts
Sample Skill (fixture)

0
installs
▌
Review It · devotts bundle
The QA front door of the DevOtts lifecycle family — plan-it plans, build-it builds, /review-it verifies. PRIMARY mission: run the unit tests, e2e tests and test-cases generated at plan phase to prove the build obeys the plan — the independent verification leg of the plan→build→review triangle. Also verifies third-party side-effects (the Airtable class), staging/prod deploys (deployed-code ladder + [REAL] re-runs), and runs a severity-tiered PR review. Routes execution to full-qa, iterate, chrome-cdp-control, make-eval and parallel-lifecycle — never re-implements them — and enforces an 11-rule gate catalog that makes false-VERIFIED claims un-shippable. Invoked with no Test Contract it never refuses and never self-grades — it runs the no-contract ladder and tags every verdict AUTHORED or DERIVED. Use when the user says "/review-it", "review it", "verify the build", "run the test contract", "QA this feature", "verify this deploy", "review this PR", or when build-it reaches its QA phase.

0
installs
▌
Pr Review · devotts
SECONDARY review-it mode — the PR-review PROCESS wrapper. Produces severity-tiered (BLOCKER/MAJOR/MINOR/NIT), evidence-cited (file:line) findings with a blocking-vs-advisory split, loading the consumer project's checklist from .claude/review-config.md (works with zero config via a generic checklist). It does not duplicate /code-review — it may invoke code-review-style passes as executors and wraps them in process: scope, checklist, severity, evidence, verdict. Invoked by /review-it when the target is a PR ref/branch/diff, or standalone when the user says "review this PR", "review the diff", "pre-merge review".

0
installs
▌
Side Effects · devotts
Third-party side-effect verification — the Airtable failure class, first-class. Verifies that writes to external systems (Airtable, Slack, Shopify, CRMs, payment providers) actually landed, by reading the record back from the target system's OWN surfaces — API GET for field values AND the UI render for display semantics — with read-back stability (2–3 reads) and real-target coverage. Invoked by /review-it when a DoD includes third-party writes, or standalone when the user says "verify the write landed", "check the Airtable/Slack/Shopify record", "did the integration actually save it". A write verified only by sender-side success is not verified.

0
installs
▌
Deploy Verify · devotts
Staging/prod deploy verification — proves the build actually works in the live environment, not that a dashboard says "Ready". Runs the deployed-code ladder (curl the new route, grep served HTML, digest-pin — merged ≠ deployable ≠ deployed), re-runs the Test Contract's [REAL]-tagged cases against the live env under Tier-2 rules, walks the release checklist (rollback, monitoring, feature flags, migration safety, credential hygiene), and emits READY / NOT-READY with per-item evidence. Invoked by /review-it when the target is "staging"/"prod"/a release, or standalone when the user says "verify the deploy", "is it actually live", "pre-release check", "can we ship this".

0
installs
▌
Build It · devotts bundle
Autonomous goal-to-DoD delivery orchestrator — the build stage of the *-it lifecycle ("plan-it plans it, build-it builds it"); formerly named fable-it. Hand it a goal and a numbered Definition of Done and it runs the whole job to completion, typically unattended, enforcing checkable gates (turn-end, claim, state-change, phase-boundary, delegation, and safe-parallel: interlock, worktree, integration), disk-backed run state, an evidence ledger that makes VERIFIED a lookup, a fresh-eyes verification pass, and an honest per-criterion report. Model-adaptive for Sonnet 5 and Opus 4.8; host-agnostic (Claude Code, Cursor, Codex, Copilot and any SKILL.md-compatible agent).

0
installs
▌
Full QA · devotts
Generic autonomous QA suite for any project — reads a test plan (markdown file or inline spec), runs all tests using Chrome CDP + iterate cycles, fixes bugs found, and produces a final QA report. Works with any web stack. Trigger with /full-qa [path-to-test-plan.md]. Use before releases, after significant changes, or whenever you want a full system health check.

0
installs
▌
Launch · devotts
Mission control for autonomous projects — analyzes tasks, recommends approaches (sub-agents/teams), sets up environment (features, evals, hooks, init.sh), suggests and configures tooling (MCPs), and launches the work.

0
installs
▌
Iterate · devotts
Autonomous multi-cycle problem solver for complex tasks that require diagnosis → fix → test → verify loops. Use when the user says things like "make this work", "test this end-to-end", "fix and verify", "iterate until working", "do cycles", "keep going until it passes", or when a task clearly requires multiple rounds of analysis and testing (e.g., debugging a pipeline, seeding data, verifying integrations, making an API flow work). Also activates when the user wants autonomous QA, test runs, or system-level verification. Splits heavy work across subagents to preserve context and speed up iteration.

0
installs
▌
Make Eval · devotts bundle
Build a small, deterministic evaluation harness for an LLM-backed function — especially classifiers with a closed label set (intent routing, priority tagging, category assignment, yes/no extraction). Use this whenever you have a function that calls an LLM and you need to prove it works, catch regressions, or demonstrate validation/guardrail thinking. Trigger on phrases like "eval", "test the LLM", "is the classifier reliable", "validate the model output", "confusion matrix", "how do I know the prompt works", or any time an LLM boundary needs evidence rather than vibes. Defaults to a minimal local harness (no cloud dependency). When the project already has LangSmith wired (langsmith in package.json + LANGSMITH_API_KEY in env), generates the LangSmith mode instead — upload-dataset.ts + run-eval.ts + auto-wires `pnpm eval` and `pnpm eval:upload`, and prints clickable dataset + experiment URLs on every run. Always exact-match scored, never LLM-as-judge for closed labels.

0
installs
▌
Test Coach · devotts
Real-time test/assessment coaching assistant. Reads the browser screen via CDP (Playwright on localhost:9222) and provides instant answers, guidance, and strategy while the user takes online tests, assessments, or quizzes. Trigger this skill when the user mentions taking a test, assessment, quiz, exam, or evaluation and wants real-time help — e.g. "help me with this test", "watch my screen and give me answers", "coach me through this assessment", "I'm taking the Wonderlic/PDA/Kolbe", "read my test questions". Do NOT trigger for studying or test prep without a live browser session. Do NOT trigger for automated test-taking (clicking answers) — this skill is read-only coaching, not browser automation.

0
installs
▌
Cicd Deploy · devotts
Cicd Deploy

0
installs
▌
Launch Light · devotts
Lightweight productivity bootstrap for greenfield TypeScript projects — pnpm workspace with apps/api/ (NestJS) by default, BFF-ready so adding apps/web/ later requires zero restructuring. OpenRouter + LangSmith env slots, a colored start-services.sh launcher that hard-fails when a service doesn't come up, README, and .env.example. NO Docker, NO Postgres, NO infra. Generates a runnable scaffold in seconds without asking domain questions. Use when the user says "/launch-light", "init", "init setup", "bootstrap minimal", "spin up Nest", "give me a scratch Nest app". Hands off to /launch-scratch-project the moment Docker/Postgres/Redis/queues/multi-service enter the picture.

0
installs
▌
Deploy Router · devotts bundle
Pre-deploy coordination gate for parallel Claude Code sessions — run BEFORE /engine-deploy, /deploy-brain-app, or ANY command that mutates shared deploy infrastructure (terraform on the tenants box, box tenants.json, ghcr pushes, kubectl mutations, apps-box compose). Detects other active sessions across ALL subscriptions, classifies them (DEPLOYING-NOW / ABOUT-TO-DEPLOY / PROD-DATA-WRITER / UNRELATED), negotiates via SendMessage, claims resource-scoped locks + the box-wide apply-mutex, re-reads the live tenant set, and hands the deploy a coordination context block. This skill NEVER deploys anything itself. Trigger with "/deploy-router", "route the deploy", "coordinate the deploy", "check for parallel deploys", "who else is deploying", or as step 0 whenever the user says "ship it", "deploy this", "push to prod" and more than one session may be active.

0
installs
▌
Shopify Forge · devotts
Full-stack Shopify theme developer and ecommerce execution agent. Builds, customizes, and ships Shopify themes from design files (HTML/Figma/images). Manages products, collections, publishing, and end-to-end QA with visual verification.

0
installs
▌
Iteration Impl · devotts
Implement an iteration on top of a /launch-light scaffold (or any minimal NestJS / Next.js project). Two modes auto-detected: API iteration (domain layer on Nest) and FRONTEND iteration (Next.js/Vite UI built from HTML/Figma/image mockups, consuming an existing API). Reads the problem statement, locks 2-4 ambiguities, runs a simplicity governor that aggressively cuts ceremony, AUTHORS RED tests on the load-bearing logic IN THE MAIN THREAD (not the agents — provenance matters), then fans the implementation build across 3-4 parallel subagents in one message. The RED tests are the verdict; agent "done" claims are not. Iteration-aware: detects existing state and runs deltas on v2+. Use when the user says "/iteration-impl", "build the iteration", "implement v1 / v2 / v3 of <feature>", "add the domain", "build the nextjs frontend", "implement the UI from <mockup>", or hands you a problem statement after /launch-light. Defaults skewed toward NO ceremony. Common typos handled: "/implement-impl", "/impl", "/iter-impl"

0
installs
▌
Favicon Builder · devotts
Design and ship a brand favicon that survives 16×16 — one recognizable shape, substantial strokes, and the Next.js App Router auto-detection pipeline so the asset actually appears in the browser tab. Use when replacing a default framework favicon, when the brand has a logo or wordmark too complex for 16×16, or when the user mentions "favicon", "site icon", "browser tab icon", or "the little icon next to the URL".

0
installs
▌
Vps Maintenance · devotts
Install or update a recurring VPS housekeeping pack on any Linux + Docker VPS — daily Postgres logical backups, weekly Docker prune, weekly DB retention cleanup, and a 5-min memory pressure alert to Slack or Telegram. Use when the user says "set up vps maintenance", "/vps-maintenance", "install housekeeping crons", "add backup cron", "add docker cleanup cron", "add memory alert", "set up retention cleanup", "harden this vps", or when a new VPS is being commissioned and needs the standard ops baseline.

0
installs
▌
Chrome Cdp Control · devotts
Chrome CDP Control

0
installs
▌
Cicd Deploy Builder · devotts
Cicd Deploy Builder

0
installs
▌
Landing Page Builder · devotts
Build premium landing pages that don't smell like AI slop — concept-first atmosphere, real proof points, motion that respects users, and JSON-LD that gets you cited by Google rich results and AI search engines. Use when building a marketing site or landing page from scratch (Next.js / Tailwind / shadcn), iterating on a hero that feels generic or "Bootstrap-y", launching a SaaS / B2B / agency single-page offer, or requesting dark + light hero variants for A/B feel.

0
installs

32 published skills