* Get the data directory for neural pattern persistence * Uses .claude-flow/neural in the current working directory, * falling back to home director

* Refresh a device's state from the real hardware and recalculate * its trust score.

* Get or create the singleton graph database instance

* Get the full trust history for a specific agent, ordered chronologically.

* Get last learning time * * @returns Learning time in milliseconds

Get the resolved auto memory directory path

* Get all patterns

* Get performance metrics

Get task progress

* Get prompt registry for external registration

* Produce a human-readable recommendation based on the coherence score.

Risk score for a (origin, selector) pair, computed from break-history.

* Get all workers for a session * * @param sessionId - Session ID * @returns Worker instances

* Get severity level as number

* Get web3.storage token from environment or config

* Validate a mutation through MutationGuard before executing. * Returns true if the mutation is allowed, false if rejected. * When guard is unavaila

* Handle an error and return a structured response

Check if gh CLI is available

* Initialize the WASM module

* Initialize the WASM module

* Build one ByzantineConsensus node with `peerCount` known peers (so * byzantineF() = floor(((peerCount+1)-1)/3)), then feed it `prepareSenders` * d

* Check whether a belief's confidence meets the minimum threshold for action. * * @param beliefId - The belief to check * @returns true if co

* Check if circuit is open

* Check if an error category is retryable

Restore KV cache from an RVF-compatible binary file.

* Load all migration files from the migrations directory

* Log message if verbose mode is enabled

* Measure execution time of an async function.