Agent Claude Code
Coordinator-only Teardrop research subagent for durable GPT-Researcher reports covering security audits, architecture decisions, product improvements, roadmap choices, and competitive comparisons.
Skill Claude Code
Use for complex, unfamiliar, contested, or cross-source questions needing evidence synthesis, including literature reviews and pre-implementation research; not routine lookups or narrow repository debugging. Read-only.
Skill Claude Code
Use when reviewing Teardrop Python dependencies for vulnerabilities, upgrade drift, changelog risk, and safer alternatives using the local audit script plus primary package metadata.
Skill Claude Code
Use when running a short human-in-the-loop Teardrop development cycle that needs scoped research, a concrete plan, implementation, and strict verification with high token efficiency.
Skill Claude Code
Use when upgrading, deprecating, or replacing Teardrop LLM router models, including pricingrules updates, benchmark catalogue changes, model sunset plans, and migration scaffolding for new provider releases.
Skill Claude Code
Use when the user wants query-driven, cited repository research for Teardrop security, roadmap, improvements, or competitive questions and a Git-tracked draft report.
Skill Claude Code
Use when reviewing code, research, or plans for bugs, inconsistencies, security issues, and quality.
Skill Claude Code
Use when implementing, refactoring, and writing high-quality code based on plans or research.
Skill Claude Code
Use for any work touching Teardrop billing (x402/credits/Stripe), pricing, marketplace (platform tools vs org tools), MCP gateway, A2A delegation, SSRF, or financial database migrations. Provides authoritative invariants and file map.
Skill Claude Code
Use after adding tools, changing planner behavior, adjusting pricing, or modifying model routing. Guides writing eval tasks, running baseline diffs, and enforcing regressions with Teardrop eval harness.
Skill Claude Code
Use when adding, reviewing, or editing a Teardrop tool in tools/definitions/. Defines the mandatory agent-consumer contract (usewhen, limitations, alternatives), the fail-closed enforcement test, and the scaffold workflow.
Skill Claude Code
Scaffold an LLM upgrade workflow using llm-lifecycle-manager. Usage: /upgrade-model [--role ].
MCP server Claude CodeCodexCursor +2
Teardrop MCP gateway: paid tools for autonomous agents (x402, credits). Remote server at api.teardrop.dev.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: