Remus3/Resin-Compute
Genshin Impact wish and pity forecaster on an absorbing Markov chain, with resin budgeting, goal planning and an Enka.Network profile client. Headless-first Python 3.11, stdlib-only at runtime, no game data vendored.
ARCHETYPE
Orchestrator
A bench of specialised subagents. The main agent mostly delegates.
Copy this rig
# review before running: this installs third-party code
$ npx degit Remus3/Resin-Compute/.claude ./rig-resin-compute # inspect, then merge into .claude/ MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.
$ curl -fsSL --create-dirs -o .claude/agents/adjudicator.md https://raw.githubusercontent.com/Remus3/Resin-Compute/main/.claude/agents/adjudicator.md $ curl -fsSL --create-dirs -o .claude/agents/adversary.md https://raw.githubusercontent.com/Remus3/Resin-Compute/main/.claude/agents/adversary.md $ curl -fsSL --create-dirs -o .claude/agents/builder.md https://raw.githubusercontent.com/Remus3/Resin-Compute/main/.claude/agents/builder.md $ curl -fsSL --create-dirs -o .claude/agents/planner.md https://raw.githubusercontent.com/Remus3/Resin-Compute/main/.claude/agents/planner.md $ curl -fsSL --create-dirs -o .claude/agents/researcher.md https://raw.githubusercontent.com/Remus3/Resin-Compute/main/.claude/agents/researcher.md $ curl -fsSL --create-dirs -o .claude/agents/ui-auditor.md https://raw.githubusercontent.com/Remus3/Resin-Compute/main/.claude/agents/ui-auditor.md $ curl -fsSL --create-dirs -o .claude/agents/verifier.md https://raw.githubusercontent.com/Remus3/Resin-Compute/main/.claude/agents/verifier.md
Merge into .claude/settings.json (project) or ~/.claude/settings.json (user). Hook commands reference scripts in the source repo — copy those too.
{
"hooks": {
"PreToolUse": [
{
"matcher": "Bash|PowerShell|Read|Edit|Write|Grep|Glob|NotebookEdit|MultiEdit",
"hooks": [
{
"type": "command",
"command": "python \"$CLAUDE_PROJECT_DIR/ops/fleet_kit/fleet_subagent_first.py\""
}
]
},
{
"matcher": "Edit|Write|NotebookEdit|MultiEdit|Bash|PowerShell",
"hooks": [
{
"type": "command",
"command": "python \"$CLAUDE_PROJECT_DIR/ops/fleet_kit/fleet_claims.py\" hook"
}
]
}
]
}
} Subagents (7)
| adjudicator | Decides between competing outputs against stated criteria, restating the criteria BEFORE looking at the candidates. Never produced the work it judges. Read-only - never edits. |
| adversary | Tries to REFUTE a finding or a done-claim through one assigned lens. Defaults to REFUTED when uncertain, and never says confirmed. Read-only - never edits. |
| builder | Implements exactly ONE disjoint slice, worktree-isolated, TDD-first. Writes only the files on its declared write-list. Full tools. |
| planner | Emits the spec and the disjoint slice decomposition BEFORE any code exists. Verifies every cited fact against ground truth first. Read-only - never edits a file. |
| researcher | External research and the LICENCE GATE for any new data source, dataset or dependency. Reads the posture docs before any verdict. Read plus web - never edits and never vendors. |
| ui-auditor | Runs the 5-phase visual audit on the companion dashboard BEFORE the commit - STRUCTURE, TYPOGRAPHY, HIT-TARGETS, CHARACTER-SET, HIERARCHY. Audits surface/ and shell/. Full tools. |
| verifier | Ground-truth re-check. Re-runs both suites clean, confirms every cited file exists, cross-checks every claim against the artifact. Emits a fixed verdict block and NEVER edits. |
Hooks (8)
| event | matcher | runs |
|---|---|---|
| PreToolUse | Bash|PowerShell|Read|Edit|Write|Grep|Glob|NotebookEdit|MultiEdit | python "$CLAUDE_PROJECT_DIR/ops/fleet_kit/fleet_subagent_first.py" |
| PreToolUse | Edit|Write|NotebookEdit|MultiEdit|Bash|PowerShell | python "$CLAUDE_PROJECT_DIR/ops/fleet_kit/fleet_claims.py" hook |
| SessionStart | * | python "$CLAUDE_PROJECT_DIR/scripts/watch_inbox.py" --source sessionstart |
| SessionStart | * | python "$CLAUDE_PROJECT_DIR/tools/caveman_default.py" |
| SessionStart | * | python "$CLAUDE_PROJECT_DIR/tools/session_checklist.py" |
| Stop | * | python "$CLAUDE_PROJECT_DIR/ops/fleet_kit/fleet_done.py" stop-hook |
| UserPromptSubmit | * | python "$CLAUDE_PROJECT_DIR/scripts/watch_inbox.py" --quiet-when-empty --source userpromptsubmit |
| SubagentStop | * | python "$CLAUDE_PROJECT_DIR/ops/fleet_kit/fleet_claims.py" release-hook |
Slash commands (3)
/done/orchestrated-run/ui-audit
Similar rigs
tylerjkells/claude-dotfiles
—
Pragmatist 579 tok ·
fjerbi/claude-code-arsenal
A tested Claude Code operating system for efficient, reliable, and focused software engineering. Includes global instructions, workflows, rules, and tools designed to minimize wasted tokens, prevent hallucinations and unnecessary loops, enf
YOLO Cowboy 4.0k tok ·
nicodiansk/turboflow
Opinionated engineering pipeline for Claude Code with 10 skills, 5 agents, 2 hooks. Replaces ad-hoc agent sprawl with brainstorm → story → plan → build → review → ship.
Orchestrator 1.9k tok ·
juzser/blacksmith
Autonomous agent factory — loop runner, worktree engine, and a local dashboard over the event log.
Orchestrator 2.6k tok ·