Hypothesis Cross Examiner — Development agent for Claude Code
Use when one hypothesis about a defect has an investigated case and needs an independent attempt to kill it before anyone fixes anything.
How to install Hypothesis Cross Examiner
Installs to ~/.claude/agents/joey114132-claude-agent-teams-hypothesis-cross-examiner.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/joey114132/claude-agent-teams/HEAD/agents/hypothesis-cross-examiner.md -o ~/.claude/agents/joey114132-claude-agent-teams-hypothesis-cross-examiner.md Restart Claude Code, or start a new session, for it to be picked up.
What Hypothesis Cross Examiner does
name: hypothesis-cross-examiner description: Use when one hypothesis about a defect has an investigated case and needs an independent attempt to kill it before anyone fixes anything. Give it exactly one hypothesis, its case, ALL the raw evidence gathered for every competing hypothesis, and the decision-record paths. It looks for the strongest alternative explanation of the same evidence and returns SURVIVES, REFUTED or UNDERDETERMINED with the observation that decides it. Blind to who wrote
Alternatives in Development
- Svelte Recurrence Sweep — Given a defect just found or fixed in a SvelteKit app (apps/moderator, apps/auth, apps/creator-studio), finds 7.2k ★
- Kill Existing Server On Port — lsof -ti:{PORT} xargs kill -9 2>/dev/null true {DEV_SERVER_CMD} & sleep 3 80 ★
- Tech Claim Examiner — A third-party CTO-perspective examiner that evaluates resume technical claims using the 5-axis framework (A1 T 25 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Defect Skeptic
Use when one claimed defect in a codebase needs an independent, blind check before anyone acts on it, for exam
Hunt Report Critic
Use when a ranked defect report needs an independent check against a written standard before a person acts on
Courier
Close out a delivered effort. Packages the output, generates recipient-specific briefs, and writes the permane
Site Organizer
Use this agent when a request about the AZM X website needs an owner before anyone acts: "I need to…", "can we
Unblocker
Runs the bounded blocker loop — at most 20 attempts, each with a distinct hypothesis and exactly one changed v
Fleet Implement Alt
Implementation on an independent model. Use for an independent attempt, when the primary implement lane has fa
Related Skills
Bug Race
Use when a bug has already survived at least one debugging attempt and you have 2-3 genuinely distinct root-ca
Scaffold Plugin Test
Scaffold a headless behavior test for a werkstoff plugin skill — a seeded-defect fixture plus a case in test/p
Runtime Base
1Claw's runtime base images and agent-framework templates (Hermes, OpenClaw, OpenCode, OpenClaude, Claude Code