数据与 AI技能

Plan a feature before writing any code. Use this skill at the START of non-trivial feature work — when the user says they want to "add", "build", "implement", "create", "develop", or "ship" a feature, endpoint, component, page, service, or capability. Especially trigger when the user's prompt is short or under-specified ("add a dashboard", "build an auth flow", "let's do user activity reports") — those need planning, not direct execution. The skill walks through requirements clarification, codebase exploration, sketches a Mermaid architecture/flow diagram of the proposed design, and produces a plan.md document the user can hand off to a fresh AI session for execution. Skip only when the change is genuinely trivial (typo fixes, single-line tweaks, renames). When in doubt, plan.

中文简介正在准备中。当前显示原文说明。

Guidance for Tauri v2 NFC plugin with session handling and data validation.

中文简介正在准备中。当前显示原文说明。

Guidance for Tauri v2 SQL plugin setup, migrations, and safe query access.

中文简介正在准备中。当前显示原文说明。

Guidance for Tauri v2 stronghold plugin with encrypted storage and sensitive data handling.

中文简介正在准备中。当前显示原文说明。

Generate fantasy character names in zh-TW with original spelling and etymology, across human, elf, dwarf, orc/beastfolk, and halfling naming traditions. Use when the user needs a name for a character, NPC, party member, villain, shopkeeper, or monster, asks to think of, suggest, or rename such a character, names entities for a MUD/TRPG/game world or worldbuilding document, or writes lore, quests, presets, or seed data that require in-world personal names.

中文简介正在准备中。当前显示原文说明。

Tunes noisy detections by analyzing historical alert data to quantify false-positive sources, recommend allow-list filters and threshold adjustments, and measure precision impact. Activates for requests to tune a detection, reduce false positives, or analyze alert noise and recommend filters.

中文简介正在准备中。当前显示原文说明。

Pivots on indicators of compromise across multiple data sources by correlating a seed set of IOCs against logs to find co-occurring indicators, hosts, and timeframes that expand the investigation. Activates for requests to pivot on IOCs, correlate indicators across data sources, or expand an investigation from a seed indicator set.

中文简介正在准备中。当前显示原文说明。

Hunts for unusual outbound network connections by flagging direct-to-IP traffic, uncommon ports, rare destinations, high data-egress ratios, and connections from unexpected processes in network telemetry. Activates for requests to hunt unusual outbound traffic, detect suspicious egress, or find anomalous external connections.

中文简介正在准备中。当前显示原文说明。

Detects DNS-based tunneling and data exfiltration in DNS logs: scoring high query volume to single domains, long/high-entropy subdomains, unusual record types, and encoded payload labels. Activates for requests to hunt DNS tunneling, detect DNS exfiltration, or analyze suspicious DNS query patterns.

中文简介正在准备中。当前显示原文说明。

Hunts for pre-exfiltration data staging and archiving by detecting archive-tool execution (rar, 7z, zip, tar), large compressed files written to staging directories, and collection into a single location before transfer. Activates for requests to hunt data staging, detect pre-exfil archiving, or find collection and compression before exfiltration.

中文简介正在准备中。当前显示原文说明。

Builds Zeek-based network hunting analytics by writing scripts and analyzing Zeek logs (conn, dns, http, ssl, files) to surface long connections, rare JA3s, suspicious downloads, and beaconing. Activates for requests to build Zeek analytics, write a Zeek hunting script, or analyze Zeek logs for threats.

中文简介正在准备中。当前显示原文说明。

Analyzes suspected webshells (PHP, ASPX/ASP, JSP) by detecting dynamic-execution sinks, input-to-sink data flow, and common encoder/obfuscation chains, then decoding obvious layers to reveal the command interface. Activates for requests to analyze a webshell, identify a PHP/ASPX/JSP backdoor, or deobfuscate server-side malicious scripts.

中文简介正在准备中。当前显示原文说明。

Frames a structured, testable threat-hunting hypothesis: grounding it in adversary behavior and available telemetry, defining the data sources and detection logic, and setting success criteria and outcomes. Activates for requests to start a threat hunt, write a hunt hypothesis, or plan a hypothesis-driven hunt.

中文简介正在准备中。当前显示原文说明。

Analyzes banking trojan webinject configurations to extract targeted institutions, injected JavaScript/HTML, and data-theft hooks, mapping the fraud workflow. Activates for requests to analyze banking trojan webinjects, parse a webinject config, or map targeted banks and credential-theft injections.

中文简介正在准备中。当前显示原文说明。

Give your AI agents something more useful than a prompt. Velocity through clarity.

中文简介正在准备中。当前显示原文说明。

Set up GEPA prompt optimization for an ORC workflow

中文简介正在准备中。当前显示原文说明。

Set up LLM-as-judge evaluation for ORC workflows

中文简介正在准备中。当前显示原文说明。

SQL discipline for querying Pulp Perfetto traces (.pftrace) with trace_processor — idempotent CREATE OR REPLACE PERFETTO views, GLOB not LIKE, dur = -1 incomplete-slice handling, EXTRACT_ARG for span args, joining on stable utid/upid, SPAN_JOIN PARTITIONED, and the draft→validate→execute loop. Ships Pulp's CPU/UI trace stdlib plus closed GPU startup, health, and probe views. TRIGGER when writing or debugging SQL over a .pftrace, when `pulp trace query` returns wrong/empty rows, or when the trace-analysis skill needs a query primitive.

中文简介正在准备中。当前显示原文说明。

Prove a UI or generation feature actually works before asking a human to look at it. Covers A/B-ing an imported design against its source render, driving controls headlessly so a dead button cannot ship, proving the generator really spawns, and launching the real host (VCV Rack, a DAW). Use before any "take a look" on imported UI, plugin shells, or prompt-to-artifact pipelines.

中文简介正在准备中。当前显示原文说明。

Judge whether a change made things better with demonstrable proxy measures tied to the mechanism it targets, not wall-clock time. Load before choosing acceptance measures for any before/after or A/B claim — CI and fleet speed-ups, build speed, DSP/audio quality, render fidelity. Each proxy names its data source, detection floor and sample size, and every zero is paired with a control on the same instrument.

中文简介正在准备中。当前显示原文说明。

Coordinate a cross-session or cross-machine handoff — snapshot the open work, write a status doc to the pulp-planning repo on main, and emit a ready-to-paste goal prompt that links it, so a fresh session (often on another machine) can pick up and finish. Use when the user says "coordinate a handoff", "hand this off", "move this work to a new machine", "we're retiring this session", or asks for a status doc plus a goal for someone else to continue.

中文简介正在准备中。当前显示原文说明。

Research, author, validate, render, and archive data-only Pulp Sample Heritage profiles. Use when creating or revising a sampler character profile, importing or exporting Heritage JSON, proving profile behavior with analytic and listening fixtures, recording evidence and trademark-safe provenance, or preparing a caller-owned profile artifact bundle.

中文简介正在准备中。当前显示原文说明。

Validate, install, update, list, rescan, remove, and reveal data-only Pulp content packs for installed plugins. Use for presets, themes, samples, sample banks, wavetables, and other end-user data that must be reviewed before install or update.

中文简介正在准备中。当前显示原文说明。

Convert any pile of documents (PDF, Word, HTML pages, wiki exports, PowerPoint slides) into clean, LLM-ready markdown context: inventories the sources, picks the right extractor per file type (pdftotext or PyMuPDF for PDF, pandoc or mammoth for docx, readability extraction for HTML.

中文简介正在准备中。当前显示原文说明。