Installer now creates an initial daily log so watchdogs pass on fresh workspaces

Adds install_memory_foundation.py and quick-install instructions

dr-memory-foundation 2.0.0 - Initializes meta information with the addition of the _meta.json file. - No other user-facing changes or feature updates in this release.

Added a simpler install + apply activation flow. Added Apply dr-memory-foundation to this workspace as the preferred activation phrase. Added references/APPLY.md with an idempotent apply checklist. Updated the skill docs so configure_prompt.md is now a fallback, not the primary activation path. Improved workspace-apply behavior guidance so existing memory files are preserved and normalized more safely.

- Retooled the skill from "dr-context-pipeline-v1" to "dr-memory-foundation," focusing on file-based memory structure instead of pipeline logic. - Introduced reference templates for memory dashboards, topic files, glossary, and policy catalog to support standardized memory layouts. - Removed pipeline, routing, compression, and schema files to streamline usage for memory organization. - Updated documentation to guide users on setting up and maintaining a memory folder structure with dedicated templates.

**Major update: the skill now provides a deterministic, schema-validated context pipeline for agent memory retrieval and compression.** - Replaced template-based memory folder setup with a full context pipeline: deterministic routing, retrieval, compression, linting, and regression testing. - Added retrieval/compression/compression prompt files, schemas (JSON), and routing configuration for robust, standardized pipelines. - Introduced golden regression tests for pipeline verification. - Updated documentation to reflect pipeline stages, prerequisites, and usage guidance. - Renamed the skill to clarify focus on context pipeline (previously “dr-memory-foundation”, now “dr-context-pipeline-v1”). - Removed previous dashboard/topic/glossary template files and detailed folder structure instructions.

# Release notes — dr-memory-foundation v1.0.0 ## What this is **DR Memory Foundation** is an opinionated, file-based memory layout for OpenClaw-style agents. It makes memory **easy to retrieve from, audit, and maintain**, and is designed to support retrieval-first context pipelines without turning memory into an unbounded blob. ## What’s included (templates) - `MEMORY.md` (long-term index): intended to stay **small** — preferences + pointers only (no daily logs). - `memory/always_on.md`: a tiny **policy header** plus a **topic catalog** with keywords. - Dashboards / registries: - `memory/now.md` — current priorities + key links - `memory/open-loops.md` — follow-ups / unresolved items - `memory/automation.md` — automation registry (cron jobs, schedules, delivery notes) - Topics: - `memory/topics/glossary.md` — canonical terms/definitions - Pattern for adding more: `memory/topics/<topic>.md` - Daily logs: - `memory/YYYY-MM-DD.md` — raw chronological notes (kept out of `MEMORY.md`) ## Why this structure works well - Promotes **stable topic files** instead of dumping everything into a single memory document. - Adds an explicit topic catalog so retrieval can quickly decide what to pull. - Clean separation of concerns: - “Standing truth/rules” → topics + glossary - “What matters now” → now/open-loops - “What happened” → daily logs ## Embedding/retrieval note (lite embeddings) This structure is optimized for **lightweight embedding-based retrieval** (a “lite” embedding model) by: - keeping important memory files short and well-scoped, - using a topic catalog + keywords to improve retrieval targeting, - avoiding giant monolithic memory files that make semantic search noisy.