AI & ML

morgs32 Skill

Codify an agreed convention or reusable repository-specific lesson in the active project's {root}/llm-wiki/** guidance. Use after a code change, when the user names a local pattern, or with no prompt to preserve conclusions from the current chat. Do not use for shared skills or morgs32/llm-wiki.

morgs32 Bundle

Update shared patterns or skill source in morgs32/llm-wiki and publish the coherent change through a GitHub pull request. Use from this repository when the user asks to change or publish its patterns or skills. Do not use for a consuming repository's local {root}/llm-wiki/** guidance.

otrebu-agents Bundle

Comprehensive web research via Parallel Search API. Use when user requests parallel search for deep multi-source research, technical analysis, learning new topics, current events, or comparative studies. Returns LLM-ready ranked URLs with extended excerpts (up to 30K chars). Single API call handles multiple query angles with automatic deduplication.

karstenheld3 Bundle

ONLY apply when user EXPLICITLY mentions "Playwriter". Default browser automation is ms-playwright-mcp.

karstenheld3 Bundle

Automates browser interactions via Microsoft Playwright MCP server (@playwright/mcp). Use when navigating websites, filling forms, taking screenshots, scraping web data, testing web UI, or automating any browser-based task. Covers persistent sessions, extension mode for existing browser tabs, and 40+ browser automation tools.

ddunnock Bundle

Creates plugins (skills and MCPs) by guiding through the authoring process. Use when users want to create or update (1) skills that extend Claude's capabilities with specialized knowledge and workflows, or (2) MCPs (Model Context Protocol servers) that provide tools for Claude Desktop.

msbart2 Bundle

Review a tech-talk's README and produce a deck.recipe.yml. Requires an independent cross-model Rubber Duck critique of section weighting, narrative arc, and coverage gaps. Always overwrites any existing recipe. Triggers: "review the talk", "create the recipe", "is this the best use of our time", "section weighting", "coverage gap", "recipe".

msbart2 Bundle

Review an exec-talk's README and produce an exec.recipe.yml. Requires an independent cross-model Rubber Duck critique of section weighting, executive audience fit, narrative arc, and action clarity. Always overwrites any existing recipe. Triggers: "review the talk", "create the recipe", "is this landing for execs", "section weighting", "coverage gap", "recipe", "executive framing".

msbart2 Skill

Use after an approved content refresh when recipeReview.required is true, or when the user asks to refresh an existing recipe. Preserves the current recipe as a baseline and requires an independent cross-model Rubber Duck critique before structural changes are finalized. Triggers: refresh recipe, changelog recipe review, approve recipe, update existing deck recipe.

msbart2 Skill

Refresh existing tech talks from proposal-created decisions in .github/content-routing/ledger.json, or from a named talk even when the ledger is empty. Use for implementing the routing inbox, applying ledger proposals, refreshing impacted talks, processing the content refresh queue, or closing feed-driven talk updates. Orchestrates content.refresh.yml and README updates; cross-model recipe review and full slide regen only when the plan requires them.

msbart2 Skill

Use this skill when asked to verify, validate, fact-check, or review any data entry, lore fact, character bio, episode record, quote, or show content for accuracy against canonical project sources and the live database. Triggers on: check data, verify data, fact-check, accuracy check, data accuracy, check entry, verify entry, review accuracy, is this accurate, canon check, lore check, check this data, verify this, review this entry, check for errors, validate data, is this right, is this canon.

msbart2 Skill

Use this skill when asked to verify, validate, or fact-check a lore entry, character description, seed data record, or any Breaking Bad content against the canonical reference in docs/breaking-bad-universe.md. Triggers on: check lore, verify lore, fact-check, canon check, lore accuracy, is this correct, seed data accuracy.

ddunnock Bundle

Guide RCCA/8D problem definition using 5W2H and IS/IS NOT analysis. Transforms scattered failure data into precise, measurable problem statements that bound investigation scope without embedding cause or solution. Use when defining problems for root cause analysis, writing D2 sections of 8D reports, analyzing nonconformances, investigating failures, or when user mentions problem definition, problem statement, RCCA, 8D, failure analysis, or corrective action.

langadventurellc Skill

Rewrites technical text into plain English for a human reader. Use when an agent has produced prose that reads like implementation notes, or when the user asks to "humanize", "rewrite for humans", or "make this less technical" — for PR descriptions, ticket bodies, release notes, or code documentation. Returns only the rewritten text, ready to drop in.

langadventurellc Skill

Helps transform vague change requests into clear, structured requirements through targeted conversation. Use when a user has a change in mind for the codebase but hasn't fully articulated what they need — e.g. "we should add X", "improve Y", "fix the way Z works", "let's build a feature for...". Scales question depth to the complexity of the work. Produces an in-chat requirements summary suitable for handing to a ticket tool, a planner, another skill, or directly to an implementation agent.

langadventurellc Bundle

Orchestrates Trellis issue creation using Claude Code Agent Teams. Use when asked to "create trellis issues", "create and review issues", "create verified issues", or when you want issues created by a writer teammate and automatically reviewed by a reviewer teammate with direct-message fix loops.

langadventurellc Skill

Orchestrates implementation of Trellis issues using Claude Code Agent Teams. Use when asked to "implement feature", "implement trellis issues", "execute feature with teams", "implement tasks via agent teams", or whenever an agent-teams-based implementation run is desired. The lead spawns a fresh developer/reviewer pair per leaf task; teammates coordinate review/fix cycles via direct SendMessage. Commits per wave by default; supports --no-commit, --no-docs, and --version flags. Recursive by default; unplanned non-leaf issues are skipped, never expanded.

peabody124 Skill

Use when working with camera projections, intrinsics, extrinsics, triangulation, reprojection, coordinate transforms between world/camera frames, or any code that touches camera_params dicts. Covers unit conventions (mm vs meters), the OpenCV world-to-camera extrinsic convention, and the required API functions from camera.py and camera_adapter.py.

stevengonsalvez Skill

ATC (Air Traffic Control) — the single operator-visible fleet supervisor, in exactly one mode: `lite` (no LLM: deterministic scan, known-transient retry policy, safety caps, evidence) or `full` (the persistent orchestrating brain). Full mode is a plain `ainb` session (Claude or Codex) running a generated policy, woken on a heartbeat (default 15 min) built from the LLM-free `fleet needs` read — which is itself HOOKS-PRIMARY (event-sourced): ATC learns ASK/WAIT/IDLE from the materialized `current_state` table, not by scraping panes, with a tmux fallback for non-Claude agents. `atc setup` installs the global Claude Code hooks into ~/.claude/settings.json. It auto-clears the safe/blocked sessions via the fleet verbs (confident ASK → broadcast; ERR → continue within a retry cap) and escalates the uncertain ones to the phone bridge. Supersedes the `daemon` skill for managed fleets. Use to provision / inspect / tear down an ATC instance, or to switch which controller owns the fleet.

letta-ai Skill

Guide for adding new LLM models to Letta Code. Use when the user wants to add support for a new model, needs to know valid model handles, or wants to update model-specific compatibility behavior. Covers runtime catalog sources, CI test matrices, and handle validation.

letta-ai Bundle

Read the official Letta documentation (docs.letta.com) through its cached, ETag-checked fetch route. Load before ANY docs.letta.com retrieval — answering how Letta works, what Letta (or you) can do, setting up providers, models, channels, skills, memory, schedules, permissions, self-hosting, pricing, or billing, AND looking up Letta API, Agent SDK, or Letta Code reference while writing code. Do not use fetch_webpage or web_search on docs.letta.com; this skill's helper is the docs route. Never answer Letta product questions from memory alone.

peabody124 Skill

Query and analyze clinical gait lab (marker-based motion capture) data. Use when asked about c3d files, gait lab sessions, mocap markers, MocapProjection synchronization, video-mocap temporal alignment, GaitLab-MMC trial matching, gait events, gait parameters, forceplate/EMG analog data, or comparing marker-based and markerless motion capture systems.

peabody124 Skill

Find and fix JIT/fusion-hygiene misses in JAX/Equinox code: expensive work running eager (un-jitted), under-fused (many tiny jits, or per-element Python loops calling un-jitted forwards), or missing `eqx.filter_jit` on a hot path. Use when "every step is slow", "GPU sits idle", "low GPU utilization", a model "runs eager" / "not jitted" / "per-crop" / "per-frame in a Python loop", or when you want to "fuse" ops, choose a "jit boundary", set up `filter_jit`, `static_argnums` / `static_argnames`, `donate_argnums` / buffer donation, or decide `lax.scan` vs an unrolled loop. Also use proactively before shipping a hot inference path, or when benchmarking eager-vs-jit (warm + block_until_ready). Complements `jax-memory-and-retracing` — fusion buys speed but can raise peak memory and trigger retraces; check both.

peabody124 Skill

Structured semi-formal reasoning templates for agentic code analysis. Use this skill whenever Claude is asked to reason deeply about code semantics without executing it — including: comparing two patches or code changes for equivalence, reviewing pull requests or diffs for correctness, localizing bugs or faults from failing tests, answering questions about how code behaves across a repository, verifying refactors preserve behavior, or any task where Claude must trace execution paths through multi-file codebases. Trigger on phrases like "are these patches equivalent", "review this diff", "find the bug", "why does this test fail", "does this refactor change behavior", "trace through this code", "what does this code actually do", or any code analysis task spanning multiple files where getting the reasoning right matters. Also trigger when the user asks Claude to act as a code verifier, reward model, or static analyzer. Do NOT use for simple syntax questions, formatting, or code generation tasks where execution t

stevengonsalvez Skill

Fan out a single prompt to selected claude sessions across the fleet. Use when you need to apply the same instruction (e.g. `/clear`, `git pull`, `remote-control disconnect`) to many sessions at once. Routing: tmux send-keys by default (the broker is an opt-in fallback, toggled via AINB_FLEET_TRANSPORT). Refuses to run without an explicit targeting flag (--all, --filter <regex>, or --cwd <substring>) — no implicit fan-out.

stevengonsalvez Skill

Spawn a coding-agent session correctly with `ainb run`: always into a git worktree, never bare into a plain checkout. Use whenever you are about to start a Claude/Codex/Gemini/Copilot session in a terminal. Guards the invocations that silently produce a `(broken)` workspace row, two agents sharing one working tree, a `--parent` flag that never fires, or a teardown that cancels itself and leaks the session.

stevengonsalvez Skill

Workflow-backed Jarvis control panel. Runs the deterministic `hangar` workflow with verb=needs (discover → enrich → prioritize), renders the Jarvis HUD from its render-ready cards, fires AskUserQuestion per blocked session, and routes each answer back via tmux send-keys (broker fallback only). Requires the workflow gate (CLAUDE_CODE_WORKFLOWS=1). If the gate is off, fall back to the prompt-driven `/ainb-fleet:needs` skill.

letta-ai Bundle

Identify and repair degradation in system prompt, external memory, and skills preventing you from following instructions or remembering information as well as you should.

letta-ai Skill

Find other agents on the same server. Use when the user asks about other agents, wants to migrate memory from another agent, or needs to find an agent by name or tags.

letta-ai Skill

Reference for the `letta mcp` CLI, which finds and invokes MCP tools available to this agent. A system reminder already lists your connected MCP servers and the basic search/schema/call commands; invoke this skill when you need more — browsing a server's tools, passing large or file-based arguments, tuning search, or troubleshooting missing servers, tools, and errors.

letta-ai Skill

Send messages to other agents on your server. Use when you need to communicate with, query, or delegate tasks to another agent.

letta-ai Skill

Migrate memory blocks from an existing agent to the current agent. Use when the user wants to copy or share memory from another agent, or during /init when setting up a new agent that should inherit memory from an existing one.

letta-ai Bundle

Comprehensive guide for initializing or reorganizing agent memory. Load this skill when running /init, when the user asks you to set up your memory, or when you need guidance on creating effective memory files.

pchalasani Skill

Inter-agent communication via the msg CLI. Use this when you need to send messages to other agent sessions, read incoming messages, or coordinate with other agents in tmux panes.

pchalasani Bundle

Literature-backed English technical-prose writing rules (agent-style, 21 rules). Its lane is FORMAL technical prose (papers, design docs, proposals, READMEs, commit messages). Use ONLY when the user explicitly asks for agent-style by name (e.g. "apply agent-style", "write/revise this per agent-style rules", "iterate until agent-style clean"). NOT for de-AI-ing casual or voiced text — use remove-ai-patterns for that. Do not auto-trigger for ordinary prose or documentation tasks.

pchalasani Skill

For CLI agents WITHOUT subagent support (e.g., Codex CLI). Search previous code agent sessions for specific work, decisions, or code patterns.