safemate-agents

Fundamental and quantitative analysis run in deliberately separate contexts, synthesized into trade ideas with entries and invalidation levels, then stress-tested by a risk manager with veto power. Decision support only — the desk never executes trades.

2 skills · 5 agents

Parallel competitor and keyword research merged into one content brief, long-form drafting, channel-native social variants and an SEO/editorial pass — a complete launch content package from a single prompt.

2 skills · 4 agents

From raw warehouse tables to a decision-ready insight report: ingestion and transforms, isolated data profiling, hypothesis-driven feature work, model training and a synthesized findings report.

6 agents

Conditional incident-response chain and infrastructure build pipeline in one plugin: triage, parallel log/metric investigation, root-cause synthesis, reviewed Terraform and a prevention step so the same failure never pages twice.

6 agents

Six specialists that take a feature from API contract to deployed code: backend architecture, database schema, frontend build, generated test suites, an independent review gate and CI/CD wiring — installed together as one Claude Code plugin.

6 agents

Design, build, review and conformance-test Model Context Protocol servers: tool-surface architecture, scaffolded implementation, protocol and description-quality review, and live-client integration tests behind a compliance gate.

1 skill · 5 agents

Audio in, publish package out: transcription in an isolated context, chapter and quote extraction, SEO titles, show notes and tags, with an audio quality check — the textbook use of subagent context isolation.

6 agents

Market sizing and business case built in parallel, merged into one opportunity brief that drives UX concepts, launch messaging and sales outreach — so marketing and sales ship a single consistent message on launch day.

6 agents

The canonical orchestrator-worker research architecture: a lead that decomposes the question, parallel searchers with explicit objectives and boundaries, compressed citation-tagged findings and a synthesized, source-checked report.

6 agents

Parallel read-only auditors for codebase, API surface and MCP configuration, a log-mining detective, and an independent verifier that kills false positives — producing a severity-ranked remediation report.

5 agents