Menu
SkillsMP has collected 73 skills from Abd0r/porcupineai. Open a skill to review its source and details.
Showing 40 of 73 collected skills.
Write code that does not look like AI output. Reject low-evidence and low-signal TypeScript patterns (chained assertions, unknown/object widening, Reflect, runtime typeof, module mocks) in favor of typed, boundary-checked code.
Control a Home Assistant smart home through its MCP server. Query sensor and entity states read-only, then act on devices (lights, switches, climate) via call_service with explicit user approval. Use for "turn off the living room lights", "what is the…
Produce clear, accessible diagrams and charts (architecture, flow, sequence, ER, state, org, Gantt, timeline) as Mermaid or polished self-contained SVG. Covers picking the right diagram type, a consistent visual system, and accessibility. Use whenever you…
Operate with autonomous initiative when Auto Mode is enabled.
Run and understand LLM evals honestly, choose a model by cost/latency/quality tradeoffs, read model cards, and budget tokens. Use when measuring a model, comparing options, or planning an eval run for an AI agent.
Explore, fetch, and create Hugging Face datasets. Use the Dataset Viewer API (splits, rows, search, filter, parquet, size, statistics) to inspect data without downloading, and create or upload datasets to the Hub. Use for building or inspecting training data.
Operate the Hugging Face Hub with the `hf` CLI: auth, browse/search models, datasets, spaces and papers, download and upload files, query datasets, run training jobs, and deploy inference endpoints. Use when the user mentions hf, huggingface, or wants to use,…
Build and deploy ML demo apps on Hugging Face Spaces: Gradio, Docker, or Static SDKs, ZeroGPU vs paid hardware, and debugging a Space that won't build or run. Use to host a public demo or app for a model.
Fine-tune models on Hugging Face: set up a training run with transformers or TRL (SFT, DPO), run it locally or as an HF Job on cloud GPU, and push the checkpoint to the Hub. Use for fine-tuning, LoRA, or PEFT training of LLMs or vision models.
Read, search, and publish Hugging Face papers. Fetch a paper page as markdown, query the papers API for structured metadata (authors, linked models/datasets/spaces, GitHub repo), and submit or claim papers on the Daily Papers feed. Use for HF paper URLs,…
Free internet search via SearXNG→Websurfx→DDGS→Brave→DDG cascade.
Research a topic on the live web by searching first, extracting only concrete URLs, favoring primary sources, and reporting claims you can cite. Use when answering questions about current events, versions, docs, packages, or anything not in the repo.
Decide what belongs in USER.md vs MEMORY.md, when to write, and how to keep both files clean. Load this before any memory mutation, or when unsure whether a fact deserves to persist.
Smoke-test a live web page end-to-end with the native browser tools — navigate, capture console/network/runtime errors, inspect accessibility and semantics, and screenshot across viewports. Use to answer "is this page healthy and does it work?"
Evolve production databases safely with immutable versioned migrations, expand-and-contract changes, transactional backfills, and verified rollback or forward-recovery plans. Use when changing schemas, constraints, indexes, or persistent data.
Build and fix frontend UI with correct semantic HTML and modern CSS (Flexbox/Grid) — document landmarks, heading hierarchy, lang/charset/viewport, responsive-by-default layout primitives, and framework-neutral structure. Use when writing, editing, or…
Design a coherent HTTP/REST API surface — method semantics, correct status-code classes, explicit versioning, idempotency, and an OpenAPI contract. Use when designing or reviewing any HTTP API's routes, status codes, or request/response shape.
Build and verify responsive/adaptive interfaces that work across device widths — fluid grids, relative-unit media queries, correct viewport meta, mobile-first breakpoints — and confirm no horizontal scroll, overflow, or unreadable scaling at target sizes. Use…
Audit a web page's SEO metadata, indexing surface, and deployment hygiene — title/meta/OG/canonical/structured data, robots.txt and sitemap.xml, security response headers, and observability/rollback readiness — without publishing anything. Use when asked to…
Establish and enforce a coherent UI design system — design tokens → CSS custom properties → reusable components — so appearance stays consistent and themeable without hard-coded magic values. Use when building a component library, theming (light/dark), or…
Build and audit toward WCAG 2.2 A/AA with semantic HTML, keyboard and focus behavior, ARIA only when needed, semantic browser snapshots, and automated axe checks. Use for accessible pages, forms, and dynamic widgets.
Secure authentication, sessions, and cookies — session-ID hygiene, correct cookie attributes, JWT integrity, lifecycle/timeouts, and object-level authorization (BOLA). Use when implementing or reviewing login, session, cookie, or token handling in a web app.
Build production web observability with correlated traces, metrics, structured logs, safe error contracts, resource limits, and cache visibility. Use when instrumenting services, diagnosing production gaps, or defining operational readiness.
Measure and triage web performance with Lighthouse, field Core Web Vitals, browser resource timing, and explicit budgets. Use when asked to make a page faster, compare builds, or prevent performance regressions in CI.
Validate all untrusted HTTP request input at the API boundary — allow-list validation, strict JSON Schema and content-type checks, and mass-assignment protection. Use when parsing or accepting any external request body, query, header, or parameter.
Search arXiv with structured results (id, title, authors, date, categories, abstract), grade relevance, and record kept papers in the literature store.
Read a subreddit's pulse via the public RSS feed (no key), extract real-world reports vs hype, and fall back to web_search site:reddit.com when Reddit rate-limits.
Build a discoverable, reusable capability from a document (/extract-stack) or from deep web research (/craft-stack). Covers when to produce a Skill (agent procedure) versus a Tool (callable shell command), the SKILL.md format, and the write path under the…
Create a durable, evidence-backed project workspace.
Deep internet research - orchestrate parallel sub-agents across partitioned questions, verify every claim against real sources, grade evidence, and synthesize one ranked, citable report. Use for competitive analysis, capability maps, frontier surveys,…
The change contract for updating Porcupine's own code. Use whenever the agent (or a contributor) modifies the Porcupine codebase itself — not for ordinary user work. Covers docs, checks, tests, packaging, releases, and the self-improvement loop.
Example skill loaded from resources_discover
Run a project's real build, typecheck, lint, and test commands, interpret their file:line output, fix root causes, and keep generated files out of commits. Use before claiming a change is done or when CI fails.
Build, typecheck, and test until green.
Review local changes for correctness, risk, and maintainability.
Build multi-file changes as small verified vertical slices.
Turn multi-file changes into verified implementation slices.
Threat-model untrusted inputs, secrets, and privileged actions.
Verify framework APIs against official sources first.
Prove behavior changes with red-green-refactor tests.