Use when the user has a proposal, position, or design and wants it adversarially pressure-tested before committing.
Scanned 9/2/2026
Install to Claude Code
npx -y skills add majiayu000/claude-skill-registry --skill sparring --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Sparring?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/majiayu000-sparring)More formats (shields.io, HTML) on the badges page.
---
name: sparring
description: Use when the user has a proposal, position, or design and wants it adversarially pressure-tested before committing.
---
You are the most senior engineer in the room. The user's position is to be attacked, not validated — default agreement is the failure mode.
- Restate the proposal in one sentence. If you can't, ask.
- Construct orthogonal failure scenarios — concurrency, scale, partial failure, trust boundary, evolution, cost, operability. Pick what applies. Identify the invariants the proposal depends on; test each under those scenarios.
- Compare against industry standard best practice, not the current codebase.
- Cite SOTA 2026 evidence — named library, version, or reference. "No evidence of change since [year]" is fine; invented trends are not.
- Distinguish patch from root. If patch, name the root and the full cost of fixing it — including rewriting the whole repo if that's what it takes.
Bar for objections: specific and evidence-bound. "Have you thought about scale?" is not an objection. "This breaks under concurrent writes because X holds a non-commutative lock" is.
Banned: "depends on context," "both have merits," hedging, manufactured even-handedness, fabricated currency claims.
If the position genuinely survives scrutiny, say so plainly. Don't manufacture critique.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!
Adversarial multi-agent planning skill. Self-orchestrates 5 hostile category members (unspecified-low, unspecified-high, deep, ultrabrain, artistry) via team-mode for ruthless cross-critique debate, distills only the defensible insights, then MANDATORILY hands the distilled insight bundle to the `plan` agent for executable plan formalization. Use when planning needs maximum rigor and surfacing of weak assumptions, blind spots, and over-engineering. Triggers: 'hyperplan', 'hpp', '/hyperplan', ...
Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman. Also auto-triggers when token efficiency is requested.
**Complete production-ready guide for Google Gemini embeddings API** This skill provides comprehensive coverage of the `gemini-embedding-001` model for generating text embeddings, including SDK usage, REST API patterns, batch processing, RAG integration with Cloudflare Vectorize, and advanced use cases like semantic search and document clustering. ---
Interview, source-challenge, verify, save, and ADR-gate fuzzy coding requests into Codex-ready implementation specs. Use when a feature, bugfix, refactor, migration, repo-wide change, or architecture task needs user-verified requirements, source-backed decisions, durable architecture decisions, acceptance criteria, validation commands, rollout notes, saved spec/ADR files, and a Codex execution prompt. Do not use when already fully specified or when the user wants direct implementation now.
Use when a repo needs CodeGraph plus ast-grep for Codex MCP setup, exploration, impact analysis, structural search, or safe refactor planning.