Leon Letto
Community@leonletto · Seattle, WA
Building AI systems | Deep into vLLM, AI Agents, and RAG | Writing about enterprise AI | Also: home renovation & salsa dancing
Agent Skills by Leon Letto
Showing 21 vetted skills indexed across 1 GitHub repositories.
thrum-update-project
Merge session narratives and mechanical state into Thrum's project_state file.
thrum-configure-roles
Generate and update Codex role templates in .thrum/role_templates/ for the project environment.
thrum-inbox
Identify and surface unread and recent Thrum inbox messages.
thrum-quickstart
Register an agent and start a Thrum session with role, module, and intent.
thrum-send
Send direct and broadcast messages to agents via Thrum.
thrum-restart
Save a conversation snapshot and compose a resume plan for session restart.
thrum
Route user intents to appropriate thrum-* command skills across worktrees.
thrum-team
List active agents with roles, modules, and intents from the agent registry.
thrum-project-setup
Convert a project plan into beads epics, tasks, and prompts.
thrum-reply
Send replies to specific Thrum messages while preserving the original audience.
thrum-prime
Loads a complete session briefing for AI agents in a workflow.
thrum-wait
Block until a message arrives in Thrum workflows with configurable timeouts.
orchestrate
Launch AI agents in tmux sessions and produce merge-ready reports.
thrum-overview
Consolidate identity, team presence, inbox, and sync status into one overview.
thrum-load-context
Restore saved agent context after compaction or session restart.
configure-roles
Detect environment and generate role-based preamble templates for Thrum agents.
adversarial-critique
Orchestrate adversarial critiques to resolve design forks with a defended choice.
project-setup
Convert plan files into beads epics, tasks, prompts, and worktree scaffolds.
project-philosophy
Maintain a canonical project philosophy document at .thrum/philosophy.md.
verify-against-plan
Verify plan files and implementation diffs to produce structured findings.
efficient-multi-agent-research
Partition large research workloads across parallel sub-agents and consolidate findings into a report.