ElcanoTek

Build a presentation the user can actually open — a Bento deck, one self-contained .bento.html file that is its own viewer and editor, authored offline with no Gamma, PowerPoint, or external API, and delivered as an offline-only file with no update check and…

Profile a tabular data file (CSV/TSV/JSONL) before any analysis — runs a bundled stdlib script that reports row/column counts, inferred types, null rates, cardinality, numeric min/max/mean, top categorical values, and malformed rows. Use it the moment you…

Drive a real hosted browser and hand it to the user when a page needs a human — a login form, a captcha, 2FA, or a consent wall. Use it whenever a task needs a website that has no API, or when a sign-in blocks you — you get a live-view link the user opens to…

Review a git diff methodically in fixed order — correctness (edge cases, error paths, concurrency), then security (injection, authz, secrets), then test coverage, then maintainability — and report findings ordered by severity with file:line references plus an…

Turn analysis results into a decision-ready executive report — lead with the answer and recommendation, quantify impact in business terms, cap visuals at one chart or table per point, push methodology to an appendix, use plain language, and end with explicit…

Draft release notes or a changelog entry from the git history between two refs — gather commits/PRs with the exact git commands given, group by breaking changes/features/fixes, rewrite commit-speak into user-facing language, put breaking changes and migration…

Produce a decision-ready research brief from web sources — clarify the question, plan 3–5 targeted searches, triangulate every claim across at least two independent sources, separate fact from vendor marketing, cite every claim with its URL, and close with a…

Annotated template that documents this bundle's Agent Skills format — what a SKILL.md is, the three progressive-disclosure levels, and how to bundle a script. Read it when you want to author a new skill; it is a reference, not a real capability to invoke.

Analyze a CSV, TSV, JSON, or Parquet data file with pandas — summary stats, grouping, filtering, and optional charts saved as PNG. Use when a task involves understanding, aggregating, or visualizing tabular data in the workspace or at a URL.

Clone a GitHub repository into the workspace with git (authenticated when GITHUB_TOKEN is forwarded). Use when the task involves working with code from a specific GitHub repo.

Make changes to a GitHub repository and open a pull request — clone, branch, edit, commit, push with git, then create the PR via the GitHub REST API. Use when a task asks you to fix, change, or add something in a repo and propose it for review (not just read…

Extract the text from a PDF (local file or URL) and summarize it — abstract, key points, and notable figures or numbers. Use when a task references a PDF report, paper, contract, or manual.

Research a question on the open web — search, fetch several pages, extract their readable text, and write a cited markdown brief. Use when a task needs current information, facts, or a survey of sources that aren't already in the workspace.

Profile a CSV file with the Python standard library only — row and column counts, per-column type inference, null counts, and basic numeric statistics. Use when someone attaches or points at a CSV and wants to know what is in it before analysing it, or when…

Annotated template that teaches how to write an Agent Skill for this bundle — the frontmatter rules, the three levels of progressive disclosure, and what changes when the deployment is Kubernetes. Use when authoring a new skill, or when you want a known-good…

Profile a CSV or other delimited file with the Python standard library only — row/column counts, per-column type inference (integer/float/boolean/string), null counts and fill rates, and basic stats (min/max/mean/median/stdev for numerics; distinct count and…

Annotated template that teaches how to write an Agent Skill for this bundle — the format, the frontmatter rules, and the three levels of progressive disclosure. Use when authoring a new skill, or when you want a known-good skeleton (with a bundled reference…