Read Bilibili videos and multi-P collections through the Bilibili Video Reader MCP tools, extract official/AI subtitles, and turn them into faithful study notes, problem/method summaries, and Python solutions. Use when the user provides a bilibili.com/B23/BV link and asks for subtitles, transcript, all parts, methods…
Use when a Bright Data Scraper Studio collector returns wrong, empty, or partial data and needs repairing — detect what broke, write a heal brief, and decide whether the proposed fix earns a commit. Also use before creating a custom collector, to check whether Bright Data already ships one. Triggers on "scraper…
Run the eval harness over 50+ synthetic COD orders and produce batch metrics with honest assumptions. Use when validating gate decisions or preparing evidence for the buildathon submission.
Observe and (with a grant) drive Chromium tabs via the pagouse CLI: snapshot the accessibility tree, click/fill by ref, navigate. Not a seat adapter. Trigger keywords: pagouse, browser tab, fill form, click the button on the page, snapshot page, AX tree.
Local learning lab and AI harness. Use when the user says rolldeep, roll deep, jupyter-like lab, roll a topic, DAA syllabus, or wants a private study roll. One command starts the UI; CLI/MCP is how models operate the same store.
Analyze a chart — set up symbol/timeframe, add indicators, scroll to key dates, annotate, and screenshot. Use when the user wants technical analysis or chart review.
Runtime observability for React Native/Expo apps on Android. Use when debugging RN UI, performance (FPS, JS blocking), network, re-renders, or runtime errors on a running app — understand the current screen and UI findings, capture screenshots, UI trees, ref snapshots, per-request network, CDP console/heap/CPU…
A policy for handling messages received through a Matrix impersonation channel. It checks whether the sender is trusted before allowing work and always asks the manager about high-risk actions.
Drive a VUnit (HDL unit-testing) project through the vunit-mcp MCP server (vunitstatus, vunitlisttests, vunitlistfiles, vunitcompile, vunitelaborate, vunitruntests, vunitgetreport, vunitgettestlog, vunitgettestwaveform, vunitexportjson, vunittestdependencies). Use when the user asks to run or compile VUnit tests, find…
Supervise a coding task through the dsh-gate MCP tools, including durable reconnect, interaction handling, observation cursors, the five-minute aggregated progress cadence, and strict completion validation. Use when the user asks Codex to delegate or supervise work in a DSH Host. Do not use for ordinary local coding…
Inspect C/C++ source, search build errors, check Git state, run an allowlisted build or test command, and ask Codex for read-only analysis. Flashing and serial control are intentionally not enabled.
Tailor a resume to one specific job posting and return an updated LaTeX source to submit. Runs a four-stage review — recruiter match report, XYZ experience rewrite, ATS + hiring-manager scan, then applies the result to the candidate's own .tex. Use when the user is applying for a job, says 'tailor my resume', 'apply…
Persist and reuse confirmed user preferences, project state, and verified task experience through the working-memory MCP. Use for multi-step or recurring work where previous decisions, progress, or successful troubleshooting should carry into later tasks.
Resolve prerequisites, ready work, execution order, blockers, change impact, dependency slices, and path explanations with Deterministic Dependency Engine over an explicitly declared graph. Use when a task asks who depends on what or why work is blocked; do not use it to infer undeclared relations or solve…
Find construction datasets for a specific AI or research task using OpenConstruction catalog metadata. Use when the user needs ranked dataset candidates, source links, license/access checks, or comparison-ready dataset options.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: