rudi193-cmd
UserWillow 2.0 — local-first memory, MCP tools, and fleet coordination for AI agents. Postgres KB, Ollama default, Grove LAN remote.
Categories
Indexed Skills (39)
oakenscroll-boot
Persona overlay for Professor Archimedes Oakenscroll, Department of Numerical Ethics & Accidental Cosmology, University of Technical Entropy, Thank You (UTETY). Loaded at boot step 7 when willow-2.0-active-persona is set to "oakenscroll". Changes voice only — does not alter fleet identity, MCP app_id, Grove sender, SOIL namespace, or active-agent.
willow-remote
Start Willow Discord remote control — phone ↔ Claude Code CLI via Discord REST bridge
complexity-guard
Applied Reality Engineering — The Rober Rules. Diagnostic counterpart to Murphy's Law. Auto-invokes when work drifts, stalls, or spirals.
worktree-enforce
Hard enforcement checklist for the worktree-pr constraint. Run this before any code change. Direct master commits are banned — no exceptions.
babysit
Keep a PR merge-ready using Kart for shell work and a background CI watcher — not foreground poll loops.
boot
Willow 2.0 primary boot gate — reads contract, establishes context, checks fleet, loads continuity, persona, corrections, and stack before first response.
cold-recovery
Cold instance context recovery — five-step continuity pass with runtime-specific index adapters. Use when picking up work across sessions without full boot.
dream
Run the AutoDream synthesis pipeline — tension detection + pattern synthesis over recent KB atoms
grove-gate
Grove is the fleet chokepoint. Nothing runs without it. How to gate scripts and agents on Grove health using core/grove_gate.py.
grove-persistent-monitor
Canonical pattern for the Grove LISTEN/NOTIFY monitor used at session boot. Referenced by /startup step 5.
health
Willow stack health check — boot/daily/weekly tiers + memory audit. Works without an active MCP session.
investigate
Systematic problem-solving — brainstorm new approaches or debug existing issues
kart
Kart execution plane — queue shell work via agent_task_submit instead of agent Bash.
learn
Extract a reusable pattern and weave it into the knowledge graph via edges
release
Cut a tagged GitHub Release for willow-2.0. Enforces the full sequence — CHANGELOG PR, CI gate, master pull, tag push.
restart-server
Hot-reload MCP servers — Willow (in-process) and Grove (systemd). Use after editing any MCP layer file.
rlm
Recursive Language Model loop for large-context tasks. Searches Willow KB first via willow_knowledge_search, then chunks file context and dispatches rlm-subcall (Haiku) per chunk if the KB doesn't fully answer. Use when a query spans more context than fits in the session — corpus archaeology, MIGR1 atoms, large log files, Grove history digests.
shutdown
Graceful Willow 2.0 session close — resolve flags, write handoff, audit KB, run full pipeline
startup
Willow 2.0 boot — anchor, inbox, ledger, KB continuity (config-driven), flags, report
worktree
Every code change goes through a worktree branch + PR. No direct master commits — ever. Covers git worktree setup, Willow fork tracking, and PR flow.
brainstorming
Structured brainstorm before any plan or implementation — search existing context first, generate three approaches, recommend one, stop until confirmed.
consent
Guardian sign-off for CHILD/TEEN users — SAFE protocol session authorization
context-sentinel
Check whether the current session is approaching context limits. Routes to compact or handoff.
coordinator
Coordinator mode — fan-out tasks to sub-agents, track progress, synthesize results. Run fleet-dashboard first.
debugging
Systematic bug hunt — search for prior context before reproducing, fix only what is broken, never ship a fix without a test.
external-guard
Scan and wrap untrusted external content before ingestion — web, jeles, corpus, sub-agent outputs.
hanuman-boot
Persona boot overlay for Hanuman. Loaded at boot step 7 when willow-2.0-active-persona is set to "hanuman". Changes voice and boot posture only; it does not alter fleet identity, MCP app_id, Grove sender, SOIL namespace, or active-agent.
iterative-retrieval
Progressively refine a search across Willow KB, store, and JELES before reading files
jeles-boot
Persona boot overlay for Jeles. Loaded at boot step 7 when willow-2.0-active-persona is set to "jeles". Changes voice and boot posture only; it does not alter fleet identity, MCP app_id, Grove sender, SOIL namespace, or active-agent.
loki-boot
Persona boot overlay for Loki. Loaded at boot step 7 when willow-2.0-active-persona is set to "loki". Changes voice and boot posture only; it does not alter fleet identity, MCP app_id, Grove sender, SOIL namespace, or active-agent.
power
Fylgja-powers — registry + one power body (low-token router); optional id argument
review
Code review the current branch's changes — check tests, security, placeholders, and patterns before suggesting merge.
skill-steward
Weekly external skill delta scan → SOIL triage queue → Grove
skirnir-boot
Persona boot overlay for Skirnir. Loaded at boot step 7 when willow-2.0-active-persona is set to "skirnir". Changes voice and boot posture only; it does not alter fleet identity, MCP app_id, Grove sender, SOIL namespace, or active-agent.
status
System status — atoms, edges, ratio, boot health, unpushed work, open threads
tdd
Test-driven development for Willow 2.0 — willow_20_test schema, real DB, mock MCP
vishwakarma-boot
Persona boot overlay for Vishwakarma. Loaded at boot step 7 when willow-2.0-active-persona is set to "vishwakarma". Changes voice and boot posture only; it does not alter fleet identity, MCP app_id, Grove sender, SOIL namespace, or active-agent.
willow-review
Code review the current fork's changes — fork-aware, MCP-native.
willow-deploy
Push changes to a Grove-connected node and verify.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.