Memory layers: budgeted hot context (#27) #57

christian · 2026-07-09T19:37:58Z

christian commented

2026-07-09 19:37:58 +00:00

Tiers Reginald's context (#27) so the model always sees what matters without copying ticket data into the prompt.

HOT (this module) — charter + active directives + focus snapshot, packed under a hard 2k-token budget; the prompt seed, assembled fresh each turn.
WARM — the append-only directive/event ledger + digest, summarized on demand.
COLD — gitea + sidecar via query_project. Ticket bodies/comments/detail live here and are never inlined; the model fetches them by number.

assembleHotContext(inputs, budget=2000) packs by priority: focus (tiny, always kept) → most recent active directives (each while they fit ~⅔) → charter fills the true remainder, truncated on a line boundary. It measures the fixed tail exactly and reserves for header/joiner/ellipsis, so the total never exceeds budget. Plus estimateTokens (tokenizer-free, slight over-estimate) and activeDirectives (accepted/amended, most-recent-first).

Acceptance ✅ — hot assembles under 2k even with a ~34k-token charter; nothing ticket-shaped is inlined (only numbers + titles for focus). +5 core tests; full core suite + typecheck green.

Follow-up: wire assembleHotContext into the live system prompt in main (needs charter + directives + focus at chat time) — the assembly + budget is the tested core here.

🤖 Generated with Claude Code

Tiers Reginald's context (#27) so the model always sees what matters without copying ticket data into the prompt. - **HOT** (this module) — charter + active directives + focus snapshot, packed under a hard **2k-token** budget; the prompt seed, assembled fresh each turn. - **WARM** — the append-only directive/event ledger + digest, summarized on demand. - **COLD** — gitea + sidecar via `query_project`. Ticket bodies/comments/detail live here and are **never inlined**; the model fetches them by number. `assembleHotContext(inputs, budget=2000)` packs by priority: focus (tiny, always kept) → most recent active directives (each while they fit ~⅔) → charter fills the true remainder, truncated on a line boundary. It measures the fixed tail exactly and reserves for header/joiner/ellipsis, so the total **never exceeds budget**. Plus `estimateTokens` (tokenizer-free, slight over-estimate) and `activeDirectives` (accepted/amended, most-recent-first). **Acceptance ✅** — hot assembles under 2k even with a ~34k-token charter; nothing ticket-shaped is inlined (only numbers + titles for focus). +5 core tests; full core suite + typecheck green. **Follow-up:** wire `assembleHotContext` into the live system prompt in main (needs charter + directives + focus at chat time) — the assembly + budget is the tested core here. 🤖 Generated with [Claude Code](https://claude.com/claude-code)

christian added 1 commit 2026-07-09 19:37:58 +00:00

Memory layers: budgeted hot context (#27 ) b65ee4c8ad

Reginald's context is tiered so the model always sees what matters without ever
copying ticket data into the prompt:
- HOT (this module) — charter + active directives + the focus snapshot, packed
  under a hard token budget (2k). Assembled fresh each turn as the prompt seed.
- WARM — the append-only directive/event ledger + digest, summarized on demand.
- COLD — gitea + sidecar via query_project. Ticket bodies/comments/detail live
  here and are NEVER inlined; the model fetches them by number when needed.

- `assembleHotContext(inputs, budget=2000)`: focus (tiny, always kept) → most
  recent active directives (each while they fit ~⅔) → charter fills the true
  remainder, truncated on a line boundary. Measures the fixed tail exactly and
  reserves for header/joiner/ellipsis so the total never exceeds budget.
- `estimateTokens` (tokenizer-free ~4 chars/token, slight over-estimate so a real
  tokenizer stays under), `activeDirectives` (accepted/amended, most-recent-first).

Acceptance met: hot assembles under the 2k budget even with a ~34k-token charter;
nothing ticket-shaped is inlined (only numbers + titles for focus). +5 core tests;
full core suite + typecheck green.

Follow-up: wire assembleHotContext into the live system prompt in main (needs
charter + directives + focus at chat time) — the assembly + budget is the tested core.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

christian added 1 commit 2026-07-09 21:53:53 +00:00

Merge branch 'main' into feat/memory-layers d1a4c4410c

christian merged commit 717dc7348f into main

2026-07-09 21:53:57 +00:00

christian referenced this issue from a commit

2026-07-09 21:53:58 +00:00

Merge pull request 'Memory layers: budgeted hot context (#27)' (#57) from feat/memory-layers into main

christian referenced this pull request

2026-07-09 22:42:55 +00:00

Memory layers (hot/warm/cold) #27

Sign in to join this conversation.