Standardize br bead follow-ups for PR reviews and code reviews so each bead includes exact implementation instructions, API/function signatures, and acceptance criteria.
Use when dispatching work through the Hermes gateway with /claw, especially when the task may resolve slash commands or hand off into AO worker orchestration.
Use when starting feature work that needs isolation from current workspace or before executing implementation plans - creates isolated git worktrees with smart directory — from…
Guide and workflow for Gemini 3 API (google-genai) - Usage Notes. Use when you need Gemini 3 API (google-genai) - Usage Notes.
Guide and workflow for LLM Prompt Engineering - System Instruction Best Practices. Use when you need LLM Prompt Engineering - System Instruction Best Practices.
Use after every `git push` on a PR — enumerates ALL current gate failures, fixes ALL in one local pass, and pushes once. The fix-all-before-push invariant.
Sparse conversation history triage for current repo/worktree using ~/.claude/projects and ~/.codex/sessions with strict context budgets.
Run or maintain the cmux approval worker that scans cmux terminal surfaces for approval dialogs, classifies them with `codex exec`, and sends the matching approval key.
Backup all cmux workspaces, surfaces, and their current working directories via the socket API. Produces a timestamped JSON snapshot at ~/.cmux-backups/.
Use when supervising the level-up ZFC migration loop in this repo, especially when the cleanup-first roadmap is drifting, AO workers need steering, or PR sequencing must stay…
Review video evidence by extracting frames with ffmpeg and inspecting key UI states.
Always check ~/.bashrc for credentials, API keys, passwords, and configuration values before asking user
Diagnose and fix MiniMax 401 auth errors in AO workers — the root cause is always a redacted or invalid MINIMAX_API_KEY, not the model name.
Use BEFORE writing wiki/sources, wiki/entities, or wiki/concepts. Routes through /wiki-ingest; bulk generation must also create entity+concept pages.
Steering work on $USER's Ubuntu machine (jeff-ubuntu) via SSH. Use when asked to run commands, install software, check services, or manage files on the Ubuntu box.
Guide and workflow for Firebase Production Campaign Database Access. Use when you need Firebase Production Campaign Database Access.
Set up and use the Symphony launchd daemon in repositories that include orchestration/symphony_overlay/daemon.
Use when designing or reviewing model-owned contracts, output schemas, backend guards, routing, semantic parsing, intent detection, or data fields where application code might…
API skill for mvp-site-app-dev-i6xf2p72ka-uc.a.run.app. Use when: interacting with this site's APIs, automating workflows, debugging requests.
Guide and workflow for Dice Roll Authenticity Standards. Use when you need Dice Roll Authenticity Standards.
Guide and workflow for Browser Testing with OCR Validation. Use when you need Browser Testing with OCR Validation.
Diagnose why the automation system is not autonomously driving PRs to green and merge.
Use when fixing broken automation (CI, gates, hooks). A commit is NOT a fix; verify the fix is active post-deploy; restart gateway after config.yaml changes.
Use when designing or reviewing backend adjusters, corrections, defaults, suppressions, clamps, guardrails, compatibility aliases, or fallback logic that may override or reshape…
Read ~/.hermes/agent-orchestrator.yaml BEFORE declaring an --agent <X> value unsupported. Lists plugin shorthands (wafer, minimax, agy).
Use when making any "X is enabled" / "feature X works" / "cache works" claim in a running session — requires runtime probe output from the standard harness startup path, NOT a…
Maintain Claude Code settings, hooks, and agent config formats; use when editing or validating settings.json or hooks.
Detailed procedures for hook re-fire short-circuit — sentinel files, CR incremental-mode detection, Branch A caps, stale review dismissal commands
Guide and workflow for WorldArchitect.AI Local Debugging Reference. Use when you need WorldArchitect.AI Local Debugging Reference.
Run the Dark Factory DOT pipeline runner against a goal. Slash command: /factory. Implements StrongDM's Attractor pattern as an external Python runner — .dot files are the…
Create macOS launchd cleanup agents with dry-run verification. Writes cleanup script + plist installer, verifies with --dry-run, then installs and runs live.
Document both INPUT and OUTPUT JSON schemas for LLM-driven features to prevent data flow confusion
Design modal agents that cannot be exited until explicit completion choices are selected.
Audit WorldArchitect.AI dice rolls for fairness and anomalies; use when analyzing campaign dice data.
Use when proving, rejecting, or deleting backend adjustment registry entries with real LLM evidence and independent disabled-adjustment worktrees.
skill to programmatically dismiss macOS keychain prompts and debug headless keyring access issues
Control cmux tabs, workspaces, and terminal panes via Unix socket. Use when reading terminal output, sending commands to another agent's pane, switching tabs, or monitoring coder…
Securely guide teammates through cloud configuration, secrets management, and incident triage
Audit automation job runs, logs, PR processing metrics, and AI-authored remote commits.
Use BEFORE `mv`/`rm -rf` on any directory. Verifies pwd does not match target; session cwd cannot recover once its directory is gone.
Generate comprehensive, systematic test suites for any system component using a structured test design methodology.
Use when spawning, steering, or auditing Agent Orchestrator workers, especially when the user specifies exact AO parameters such as codex, runtime, project, claim-pr, or PR…
Use whenever running an AO spawn. Honors exact --agent/--runtime/--project/--claim-pr; verifies session metadata; never silently substitutes a different worker.
Use when persisting rewards_box or any data structure to Firestore. ALL paths (streaming/polling/passthrough) must canonicalize before writing.
Use Claude Code to control Google Antigravity IDE on macOS via Peekaboo. Explicitly supports targeting either the Agent Manager window or the Editor window (worktree/editor…
Canonical 7-green PR merge criteria, PR status check pattern, PR freeze discipline, and admin merge protocol
Why ~/.local/share/worldarchitect-runners/ is a runtime mirror, not the source — and how to sync changes correctly.
Use BEFORE claiming a UI bug fix is complete. Enforces RED-before-GREEN: reproduce the bug, verify the broken element, disclose environment gaps.
Define evidence requirements for end-to-end streaming chunk delivery and timing.
Measure zero-touch and one-touch rates from merged PRs — run in evolve loop cycles
Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies — from jleechanorg/claude-commands
Pre-report gate — verifies all planned evidence artifacts exist before writing comparison/validation reports
Use when asked to analyze a video file and add visual description captions or text overlays burned into the output video
Use when editing GitHub Actions in private repos. Requires self-hosted runners; flags ubuntu-latest as policy drift unless explicitly approved.
Canonical definition of zero-touch-by-operator for PR autonomy measurement. Use when evaluating, reporting, or labeling PR autonomy status.
Diagnose and fix AO lifecycle-worker backfill failures (stale worktrees, branch conflicts, claim_failed loops)
Use thinclaw MCP server — thin inference-less bridge to OpenClaw Gateway for executing tools
Override the model an `ao spawn` worker uses (e.g. force claude-sonnet-4-6 instead of the project default of wafer.ai/GLM-5.1) WITHOUT mutating ~/.hermes/agent-orchestrator.yaml.
Autonomous improvement loop for AO + Antigravity — observe, measure, diagnose, fix, repeat. Targets zero-touch rate as quality metric.
Use when installing/modifying any ~/Library/LaunchAgents plist. Requires template in repo that owns the job; @HOME@ placeholders; install-launchagents.sh cleanup.