Bellman
Community@yeachan-heo · Seoul
Dedicated Algorithmic Trader Leader of Quant.start() - Korea's biggest Quant Trading Community
Agent Skills by Bellman
Showing 85 vetted skills indexed across 4 GitHub repositories.
cancel
Cancels active OMX workflow modes and performs documented state cleanup.
design
Creates and maintains a repo-local DESIGN.md as the source of truth for product and UI decisions.
skill
Manages local SKILL.md files through list, add, edit, remove, search, and validate commands.
plan
Generates evidence-backed implementation plans with optional Socratic interview and review modes.
ultraqa
Runs adversarial dynamic end-to-end QA cycles with hostile scenarios, evidence capture, and cleanup.
doctor
Diagnose and repair oh-my-codex installation, hook, and plugin cache issues.
code-review
Reviews code diffs with parallel code-reviewer and architect agents producing severity-rated findings.
visual-ralph
Implements frontend UI from approved visual references using iterative verdict scoring and pixel-diff evidence.
best-practice-research
Gathers cited best-practice evidence from official and upstream sources for technology decisions.
analyze
Performs read-only repository analysis and returns ranked explanations with evidence citations.
worker
Executes team worker protocol with ACK, mailbox, and task lifecycle for tmux-based OMX teams.
performance-goal
Runs an evaluator-gated performance optimization workflow over Codex goal mode with durable OMX artifacts.
autopilot
Orchestrates autonomous delivery through deep-interview, ralplan, and ultragoal phases.
configure-notifications
Configure OMX notification providers for Discord, Telegram, Slack, and OpenClaw.
hud
Displays and configures the OMX two-layer statusline showing orchestration state and presets.
pipeline
Redirects deprecated $pipeline invocations to $plan and $team workflows.
ultrawork
Redirects deprecated $ultrawork parallel execution workflows to the $team skill.
ralph
Redirects deprecated $ralph workflow invocations to the $ultragoal command.
autoresearch-goal
Redirects deprecated autoresearch-goal invocations to the canonical autoresearch workflow.
ultragoal
Persists multi-goal plans and ledger events under .omc/ultragoal for Claude Code sessions.
autoresearch
Runs a stateful single-mission improvement loop with evaluator-driven iterations and durable decision logs.
ask
Routes prompts to local Claude, Codex, Gemini, Antigravity, Grok, or Cursor CLIs and persists responses as artifacts.
verify
Validates that code changes actually work through tests, builds, and direct checks.
deep-interview
Conducts Socratic requirement interviews with weighted ambiguity scoring before gated execution approval.