- run-selected-prompt.mjs: set SCHEDULER_MEMORY_FILE env var before spawning each agent so
store.mjs has a well-known path to read from (memory-updates/<ts>__<agent>.md)
- retrieve.mjs: format memories.md using m.summary + date/tags in readable markdown sections
instead of raw m.content blobs; agents can now scan injected context at a glance
- store.mjs: remove semantically-empty fallback placeholder events that polluted the store
and degraded retrieval quality; also fix default cadence 'weekly' -> 'daily' to match retrieve
- scheduler-flow.md: replace vague "write insights" instruction with a concrete three-section
template (Key findings / Patterns / Warnings) and clarify that $SCHEDULER_MEMORY_FILE is
set by the scheduler automatically
Co-Authored-By: Claude Sonnet 4.6 <[email protected]>
- Replaced fragile inline Node.js scripts in `torch-config.json` with dedicated modules: `scripts/memory/retrieve.mjs` and `scripts/memory/store.mjs`.
- Updated `src/ops.mjs` initialization logic to dynamically configure `scheduler.memoryPolicyByCadence` with correct paths based on the installation directory.
- Added default memory policy configuration to `torch-config.example.json` to guide users.
- Injected `torch:memory:list` and `torch:memory:inspect` convenience scripts into the host `package.json` during `init`.
- Removed legacy shell scripts (`run-memory-retrieval.sh`, `run-memory-store.sh`).
- Verified memory scripts function correctly and pass linting.
Co-authored-by: PR0M3TH3AN <[email protected]>
thePR0M3TH3AN