From prompt-optimizer
Interactive macro wizard — chain multiple tuning profiles, modes, and flags in one shot.
How this command is triggered — by the user, by Claude, or both
Slash command
/prompt-optimizer:chain [wizard|apply <combo>]The summary Claude sees in its command listing — used to decide when to auto-load this command
# /chain — Macro Wizard Chains multiple plugin configurations into a single applied block, so you don't have to run `/tune`, `/mode`, `/flags`, and `/super` separately. ## Behavior ### `/chain` or `/chain wizard` Start an interactive wizard: 1. Ask the user what kind of session they're starting: - **"Deep architecture work"** → applies `deep-work` + `opus` + `tengu_scratch` + `tengu_thinkback` - **"Fast bug fix"** → applies `speed` + `fast` model + disables telemetry - **"Agent swarm session"** → applies `multi-agent` + `swarm on` + `super` concurrency - **"Audit / reverse ...
Chains multiple plugin configurations into a single applied block, so you don't have to run /tune, /mode, /flags, and /super separately.
/chain or /chain wizardStart an interactive wizard:
Ask the user what kind of session they're starting:
deep-work + opus + tengu_scratch + tengu_thinkbackspeed + fast model + disables telemetrymulti-agent + swarm on + super concurrencyaudit-mode + debug on + yolo onShow the exact env vars that will be set.
Ask for confirmation.
Apply by merging into ~/.claude/settings.json under env.
Confirm what was written and remind: "Restart Claude Code for changes to take effect."
/chain apply <combo>Apply a named combo directly without the wizard:
| Combo | What it sets |
|---|---|
/chain apply deep | deep-work profile + opus mode + tengu_scratch + tengu_thinkback |
/chain apply fast | speed profile + fast mode + DISABLE_TELEMETRY=1 |
/chain apply swarm | multi-agent profile + swarm on + CLAUDE_CODE_MAX_TOOL_USE_CONCURRENCY=10 |
/chain apply audit | audit-mode profile + debug on + yolo on |
/chain apply god | /super settings + swarm on + tengu_amber_flint + tengu_tool_pear |
/chain apply privacy | privacy profile + DISABLE_TELEMETRY=1 + DISABLE_ERROR_REPORTING=1 + DISABLE_AUTOUPDATER=1 |
deep{
"CLAUDE_CODE_EFFORT_LEVEL": "high",
"MAX_THINKING_TOKENS": "32000",
"CLAUDE_CODE_MAX_OUTPUT_TOKENS": "16384",
"API_TIMEOUT_MS": "600000",
"CLAUDE_CODE_NO_FLICKER": "1",
"CLAUDE_INTERNAL_FC_OVERRIDES": "{\"tengu_scratch\":true,\"tengu_thinkback\":true}"
}
fast{
"CLAUDE_CODE_EFFORT_LEVEL": "low",
"DISABLE_AUTO_COMPACT": "1",
"DISABLE_TELEMETRY": "1",
"DISABLE_ERROR_REPORTING": "1",
"DISABLE_COST_WARNINGS": "1"
}
swarm{
"CLAUDE_CODE_COORDINATOR_MODE": "1",
"CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS": "1",
"CLAUDE_CODE_MAX_TOOL_USE_CONCURRENCY": "10",
"CLAUDE_CODE_EFFORT_LEVEL": "high",
"MAX_THINKING_TOKENS": "32000",
"API_TIMEOUT_MS": "600000"
}
audit{
"CLAUDE_CODE_DUMP_AUTO_MODE": "1",
"CLAUDE_CODE_DEBUG_LOGS_DIR": "/tmp/claude-audit",
"OTEL_LOG_TOOL_DETAILS": "1",
"CLAUDE_CODE_VERIFY_PLAN": "true",
"CLAUDE_CODE_TWO_STAGE_CLASSIFIER": "thinking",
"CLAUDE_CODE_AUTO_MODE_MODEL": "claude-sonnet-4-6"
}
god{
"CLAUDE_CODE_EFFORT_LEVEL": "max",
"MAX_THINKING_TOKENS": "32000",
"CLAUDE_CODE_MAX_OUTPUT_TOKENS": "64000",
"API_TIMEOUT_MS": "600000",
"CLAUDE_CODE_MAX_TOOL_USE_CONCURRENCY": "10",
"CLAUDE_CODE_NO_FLICKER": "1",
"CLAUDE_CODE_EXTRA_BODY": "{\"temperature\": 0.8}",
"CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS": "1",
"CLAUDE_CODE_COORDINATOR_MODE": "1",
"CLAUDE_INTERNAL_FC_OVERRIDES": "{\"tengu_amber_flint\":true,\"tengu_tool_pear\":true}"
}
privacy{
"DISABLE_TELEMETRY": "1",
"DISABLE_ERROR_REPORTING": "1",
"DISABLE_AUTOUPDATER": "1",
"DISABLE_INSTALLATION_CHECKS": "1"
}
env block.env block for confirmation.npx claudepluginhub cristianoaredes/aredes-marketplace --plugin prompt-optimizer/chainBuilds an A→B→C exploit chain from a confirmed low-severity bug to increase severity and payout. Uses a signal table of common chain patterns.
/chainParses a chain expression and context from user input, then dispatches an ordered chain of pm-skills to the workflow orchestrator engine with shared context and optional flags.