mcp-genkit-flows-skill

Memoization caches flow results based on inputs, eliminating redundant computations. This Skill executes Genkit flows with SHA-256-based cache keys and Firestore Redis storage, achieving 70-90% token savings and 90%+ cache hit rates for repeated or similar requests.

Yes. This Skill provides list_flows to discover available flows, get_flow to retrieve flow details, and execute_flow to run flows with specified inputs. It supports common use cases like generating responses, creating cover letters, and analyzing resumes.

Memoization stores computed results indexed by input hashes, serving cached outputs for identical or matching requests within the TTL window. Sub-second latency and 90%+ hit rates eliminate redundant AI computations across multi-step workflows.

The cache_stats operation provides visibility into cache performance, including hit rates and storage metrics. An index operation surfaces the full registry, enabling tracking of memoized results and cache utilization across your workflows.

Use memoization when workflows process repeated or similar inputs—batch operations, resume analysis, or multi-step orchestrations. It's most effective for deterministic flows where token savings and latency reduction outweigh cache storage costs.