crucible

Crucible

Context: $ARGUMENTS

Find everything wrong with this idea. Socratic questioning, hostile review, and adversarial simulation.

Workflow

1. Initialize (or pick up existing graph)

Check if a graph already exists from a prior steelman:

uv run ${CLAUDE_PLUGIN_ROOT}/scripts/argument-graph.py show

If empty or no prior state, build the graph from scratch. Identify the idea's core claims and assumptions and register them:

uv run ${CLAUDE_PLUGIN_ROOT}/scripts/argument-graph.py reset
uv run ${CLAUDE_PLUGIN_ROOT}/scripts/argument-graph.py add "<claim-id>" claim --obs "<fact>"
uv run ${CLAUDE_PLUGIN_ROOT}/scripts/argument-graph.py add "<assumption-id>" assumption --obs "<fact>"
uv run ${CLAUDE_PLUGIN_ROOT}/scripts/argument-graph.py relate "<claim-id>" assumes "<assumption-id>"

If a graph already exists, use it directly.

2. Socratic Elenchus (probe assumptions)

Read ${CLAUDE_SKILL_DIR}/references/socratic-method.md.

Apply Vlastos' steps:

1. Identify the thesis (what the idea claims)
2. Identify premises the proponent would agree with
3. Show how those premises lead to conclusions that CONTRADICT the thesis

Ask 3-5 penetrating questions. Each question targets a specific assumption. Present the questions to the user.

3. Structural analysis

Run the graph analysis to find gaps the questions should have exposed:

uv run ${CLAUDE_PLUGIN_ROOT}/scripts/argument-graph.py analyze

This reveals: ungrounded assumptions (no evidence), unstable claims (depend on contradicted assumptions), exposed weaknesses (no mitigation), contested claims (multiple counter-arguments), orphaned entities (disconnected from the argument).

4. Hostile review

Read ${CLAUDE_SKILL_DIR}/references/crucible.md.

For each stakeholder who would need to approve or fund this idea, ask: what is their harshest possible question? Can the idea survive it as stated?

Register weaknesses found:

uv run ${CLAUDE_PLUGIN_ROOT}/scripts/argument-graph.py add "<weakness-id>" weakness --obs "<what makes it weak>"
uv run ${CLAUDE_PLUGIN_ROOT}/scripts/argument-graph.py relate "<assumption-id>" if_fails "<weakness-id>"

5. Red team (failure modes)

Read ${CLAUDE_SKILL_DIR}/references/red-team.md.

Identify 3-5 ways this idea fails. For each failure mode:

• What happens?
• How likely is it?
• How bad is it?
• Is there a defense?

If mitigations exist, register them:

uv run ${CLAUDE_PLUGIN_ROOT}/scripts/argument-graph.py add "<mitigation-id>" mitigation --obs "<how it helps>"
uv run ${CLAUDE_PLUGIN_ROOT}/scripts/argument-graph.py relate "<mitigation-id>" mitigates "<weakness-id>"

6. Present findings

Run analysis again to see the updated picture:

uv run ${CLAUDE_PLUGIN_ROOT}/scripts/argument-graph.py analyze

Output format

Present three sections:

Socratic Questions: 3-5 questions and the assumptions they target Weakest Points: What a hostile reviewer would attack (from structural analysis + review) Failure Modes: How this fails, likelihood, severity, and whether defenses exist

Resources

• ${CLAUDE_SKILL_DIR}/references/socratic-method.md — load always. Vlastos' elenchus steps, the maieutic method, Boghossian's warning about humiliation vs. productive discomfort.
• ${CLAUDE_SKILL_DIR}/references/crucible.md — load always. The stress-test review board method, NASA/Pentagon origin, the principle that the review must be MORE adversarial than the real presentation.
• ${CLAUDE_SKILL_DIR}/references/red-team.md — load always. RAND Corporation origins, Israel's Ipcha Mistabra, groupthink as the enemy, reconnaissance before attack.
• ${CLAUDE_PLUGIN_ROOT}/scripts/argument-graph.py — run throughout to register weaknesses, mitigations, and analyze structural gaps.

Gotchas

• Run analyze before writing findings. The graph catches things intuition misses: assumptions with zero supporting evidence, claims that depend on contradicted assumptions.
• Nemeth's warning applies here too. Generic "but what if it fails?" is worthless. Every objection must cite a specific mechanism of failure.
• If a graph already exists from a prior /steelman, use it. Don't rebuild from scratch. The crucible builds on what steelman established.
• Graph entity names: short kebab-case identifiers. Observations carry the detail.
• The graph state persists in /tmp. If /verdict runs after this, it reads the same graph.