semantic-memory

Use claim-ledger to back material agent assertions with claim/evidence/provenance receipts. Create claims from facts, add evidence, verify integrity, and export bundles for audit.

Use context-governor for receipt-backed context compaction before risky handoffs or compaction events. Preserve high-risk context, store exact fallback receipts, and expand when omitted text matters.

Use when semantic-memory evidence must support an assertion, historical answer, correction, compaction, or action. Preserve witnessed retrieval, temporal scope, authority and native receipts. Applies to the RecursiveIntell memory stack; does not authorize memory maintenance or agent execution.

Explore the semantic-memory knowledge graph. Use when the user asks what is related to a topic, how two things or facts connect, what is known across the graph, or wants clusters, communities, graph paths, second-order retrieval, or an interactive graph view.

Use semantic-memory-mcp LLM output parser tools to robustly parse JSON, string lists, choices, and numbers from raw LLM output without an additional LLM call. Handles think blocks, markdown fences, and malformed JSON.

Persist durable sourced facts to semantic memory when the user asks to remember a decision, preference, correction, or verified project fact. Excludes transient progress, secrets and unverified claims.

Use semantic-memory conversation tools. Trigger when the user asks to recall past sessions, search prior conversations, list memory sessions, save a notable exchange, or log conversation messages into semantic-memory.

Audit and improve the semantic memory store. Use when the user asks to review, clean, audit, prune, maintain, reconcile, find duplicates or contradictions, inspect stale facts, or check memory health. Produces a read-only report first, then only changes memory after approval.

Ingest or refresh a repository into semantic-memory for Codex using the deterministic codebase ingester, including dry runs, namespaces, dedupe, manifests, components, and graph edges. Prefer memory-sync for routine stale repo refreshes.

Memory specialist workflow for heavy semantic-memory work. Use for multi-step memory audits, deep graph exploration, bulk namespace recall, contradiction reconciliation, conversation recall, or when memory work should be isolated from the main task.

Maintain semantic-memory quality for Codex by auditing duplicates, stale or contradictory facts, provenance, graph relationships, lifecycle output, topology, recall ranking, and direct namespace reads. Prefer memory-curator for full audit and reconciliation.

Diagnose and repair semantic-memory Codex setup, including plugin install state, MCP binary resolution, memory directory, hooks.json, config.toml, tool approvals, and smoke tests.

Use Codex subagents with semantic-memory for large repo exploration, memory audits, ingestion planning, graph traversal, conversation recall, and parallel recall while preserving verification discipline. Prefer memory-keeper for heavy memory-only work.

Keep a codebase's semantic memory current. Use when the user asks to sync, refresh, re-ingest, or update memory for a repository after pulling changes, finishing work, or finding stale project memory.

Run cargo fmt/clippy/test and store gate receipts before claiming work done.

Use local-first semantic-memory-mcp as persistent Codex memory, including setup, recall, direct fact reads, namespace browsing, fact persistence, graph-aware project memory, conversation-memory search, curation, and deterministic codebase ingestion.