jeromeetienne

Make a new release in CHANGELOG.md: move the Unreleased entries into a dated version section and regenerate the comparison links. Use this when the user wants to make a release, bump the version, or prepare release notes. Triggers on: "make a release", "release version X", "bump the version", "prepare a release". All AI steps run on the current Claude session — no OpenAI, no API key.

Create a new CHANGELOG.md in the Keep a Changelog 1.1.0 format. Use this when the user wants to add a changelog to a project that does not have one yet, or asks to start tracking changes. Triggers on: "create a changelog", "add a CHANGELOG.md", "start a changelog for this project", "set up a changelog". If a CHANGELOG.md already exists, this skill stops and points at the sibling changelog-update skill instead. All AI steps run on the current Claude session — no OpenAI, no API key.

Add entries to the Unreleased section of an existing CHANGELOG.md, written in human-readable Keep a Changelog style rather than raw commit-log dumps. Use this when the user wants to update the changelog, log a change they just made, or catch the changelog up with recent commits. Triggers on: "update the changelog", "add this to the changelog", "log this change", "update CHANGELOG.md". All AI steps run on the current Claude session — no OpenAI, no API key.

Find job ads on the web and match them against a resume (ResumeJson). Use this when the user wants to search for jobs that fit their resume, capture a specific job posting from a URL into structured data, or check how well their resume matches a role. Triggers on: "find jobs for my resume", "search for jobs", "job postings for me", "does my resume fit this job", "match my resume to this posting", "grab this job listing", or any reference to job ads / job search / job matching. All steps run on the current Claude session using your WebSearch / WebFetch tools — no OpenAI, no API key.

Convert resumes between PDF, Markdown, and the canonical JSON Resume format (ResumeJson). Use this whenever the user wants to extract a resume from a PDF into structured JSON, turn a Markdown resume into JSON, or render a resume JSON back to Markdown. Triggers on: "extract resume from PDF", "parse my resume", "resume to JSON", "convert resume to markdown", "resume from markdown", or any reference to ResumeJson / JSON Resume. All AI steps run on the current Claude session — no OpenAI, no API key.

Score, review, and optimize a resume for Applicant Tracking Systems (ATS) using the canonical ResumeJson format. Use this when the user wants an ATS score, an ATS compliance review, ATS screening questions, or a rewritten resume optimized for ATS. Triggers on: "ATS score", "score my resume", "ATS review", "is my resume ATS-friendly", "optimize my resume for ATS", "improve my resume", "ATS questions", "screening questions". All AI steps run on the current Claude session — no OpenAI, no API key. Produce a ResumeJson first with the resume-convert skill if you only have a PDF/Markdown.

Benchmark a program in a controlled environment and attribute the cost with the code graph. Use when asked to profile under a CPU/memory cap, measure server capacity ("how many requests per second before p99/latency breaks an SLO"), decide "when do I need more servers", find where runtime is spent under load, or run a workload in a constrained "one box" container. Two modes: cpu-profile (where the time goes) and loadtest (how much load it takes). Requires the codespine CLI; Docker is optional (falls back to running on the host).

Query a TypeScript codebase as a knowledge graph to answer impact, caller, dependency, and dead-code questions. Use this instead of grep/glob for "who calls X", "what breaks if I change X" (blast radius), "is X dead code", "what references this type", and "what is X connected to". Requires the codespine CLI and a built graph database.

Maintain the Open Knowledge Format (OKF) knowledge bundle under .okf/. Use whenever the user wants to update, refresh, or regenerate OKF docs, set up an OKF bundle, check OKF conformance, or when source code that the bundle documents (config schemas, web API routes, CLI commands, runtime subsystems, example crews, ADRs) has changed and the docs need to catch up. Three modes: scaffold (create the bundle), refresh (regenerate a folder's docs from current source), and check (conformance and dead-link lint). Prefer this skill over hand-editing .okf/ so the format, the folder<->source mapping, and link integrity stay consistent.

Read-only browser for any Open Knowledge Format (OKF) knowledge bundle — a tree of markdown concept files. Use it whenever the user wants to browse, explore, map, or navigate a knowledge bundle; asks "what concepts are in <bundle>", "what links to X" / "what does X link to", "show me the OKF graph", "find orphan or broken concepts", or "summarize recent changes"; or points at a directory of markdown docs with frontmatter and `index.md` / `log.md` files — even when they never say "OKF". Prefer this skill for any read-only question about such a bundle's structure, concepts, links, or history. It never edits the bundle.

Generate a short narrated video from a prompt using Remotion. The video topic is provided in the first prompt. Produces an MP4, per-scene MP3 voice-over (macOS `say` + ffmpeg), and a multi-page slides PDF (one still per scene via ImageMagick). Trigger when the user asks to "generate/render/build a video", "make a video from a prompt", "render with Remotion", "create voice-over", or to produce slides/PDF from the video scenes.

Generate speech from text (text-to-speech / TTS) and transcribe audio to text using the local voicebox API through the `voicebox-cli` command-line tool. Use whenever the user wants to synthesize a voice, read text aloud, turn text into an audio / MP3 / WAV file, pick or clone a voice, manage voice profiles, or transcribe a recording to text. Requires a running voicebox server (default http://127.0.0.1:17493).

Maintain the Open Knowledge Format (OKF) knowledge bundle under okf/. Use whenever the user wants to update, refresh, or regenerate OKF docs, set up an OKF bundle, check OKF conformance, or when source code that the bundle documents (config schemas, web API routes, CLI commands, runtime subsystems, example crews, ADRs) has changed and the docs need to catch up. Three modes: scaffold (create the bundle), refresh (regenerate a folder's docs from current source), and check (conformance and dead-link lint). Prefer this skill over hand-editing okf/ so the format, the folder<->source mapping, and link integrity stay consistent.

Convert resumes between PDF, JSON, and Markdown, and optimize them for Applicant Tracking Systems (ATS) using AI. Use this skill whenever the user wants to extract a resume from a PDF, render a resume back to PDF or Markdown, score how ATS-friendly a resume is, generate or answer ATS-style screening questions, or rewrite a resume to improve its ATS performance. Triggers on: "extract resume from PDF", "convert resume to PDF", "convert resume to markdown", "score my resume", "ATS score", "ATS review", "optimize my resume for ATS", "ATS screening questions", "answer ATS questions", "improve my resume", or any reference to `resumejson_cli`.

Search the web and do research using Perplexity AI. Use this skill whenever the user wants to search the web, look up current information, do research, or find facts online. Supports filtering by domain, country, and language, as well as multi-query batch search. For in-depth research reports use deep-search with a preset. Triggers on: "search the web", "look up", "find info on", "research", "what is X", "latest news on X", "search for X", or any request needing up-to-date information from the internet.

Record and maintain Architecture Decision Records (ADRs) in Nygard format under docs/ADRs. Use this whenever the user wants to document an architecture or design decision, asks to "write an ADR", "record a decision", "document why we chose X", set up an ADR log, or backfill/catalog the decisions already baked into an existing codebase. Also use it when reusing this mechanism in another project. Three modes: scaffold the docs/ADRs directory, "new" (interview the user to capture one fresh decision), and "backfill" (mine existing docs and code to reconstruct past decisions). Prefer this skill over writing ad-hoc markdown so numbering, the index, and the Nygard structure stay consistent.

Generate a "build in public" narrated video using Remotion. The video topic is provided in the first prompt. Produces an MP4, per-scene MP3 voice-over (macOS `say` + ffmpeg), and a multi-page slides PDF (one still per scene via ImageMagick). Trigger when the user asks to "generate/render/build a video", "make a build-in-public video", "render with Remotion", "create voice-over", or to produce slides/PDF from the video scenes.