star-chamber

Star-Chamber: Multi-LLM Craftsmanship Council

This skill is for explicit /star-chamber invocations with live progress in the main conversation. It supports --debate for multi-round deliberation. For automatic invocation on architectural decisions, the star-chamber agent handles that separately.

Advisory skill that fans out code reviews and design questions to multiple LLM providers (Claude, OpenAI, Gemini, etc.) and aggregates their feedback into consensus recommendations.

Arguments

Flag	Description	Manual Only
--provider <name>	LLM provider to use (repeatable, e.g., --provider openai --provider gemini). Defaults to all in config.	No
--file <path>	Specify file to review (repeatable). Defaults to recent git changes.	No
--timeout <seconds>	Timeout per provider request (overrides config timeout_seconds).	No
--debate	Enable debate mode: multiple rounds with summarization between rounds.	Yes
--rounds N	Number of debate rounds (default: 2, requires --debate).	Yes

Manual-only flags are skill invocation parameters interpreted by Claude Code, NOT flags passed to the star-chamber CLI. Debate mode is orchestrated by Claude Code (see Debate Mode in the protocol).

Path Setup

The skill loader provides the base directory in the header: Base directory for this skill: <path>. Set the path variable used throughout the protocol:

STAR_CHAMBER_PATH="<base directory from header>"
# e.g., STAR_CHAMBER_PATH="$HOME/.claude/plugins/cache/pragma/skills/star-chamber"

Protocol

Read and follow the full protocol from $STAR_CHAMBER_PATH/PROTOCOL.md. It contains Steps 0-4: prerequisite checks, invocation mode detection (code review vs design question), review target identification, context gathering, CLI invocation, and result presentation. Debate mode is documented as a separate section.

Auto-Invocation

This skill is not auto-invoked (model-invocable: false). The star-chamber custom subagent handles auto-invocation based on its description. This skill is for explicit /star-chamber invocations only.