leetcrypt 9b85255d80 feat(ai): backfill context on join + token-budget window ... The server already ships the full RAM message backlog in the init frame; the agent was discarding it. _seed_transcript now decrypts that history with the room key (skipping our own lines, control frames, and undecryptable blobs) so the agent has context the moment it joins instead of starting amnesiac. _window() replaces the fixed last-12 slice on both the answer and sandbox paths: it walks newest-to-oldest and keeps messages up to --token-budget (approx, ~4 chars/token), still capped at --context-window count. Keeps small local models inside their effective context. Nothing touches disk. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>		2026-06-02 17:43:02 -07:00
..
agent	feat(ai): backfill context on join + token-budget window	2026-06-02 17:43:02 -07:00
client	feat(agent): model-agnostic AI agent bridge (PoC) + pin lets-hack demo to main	2026-06-01 02:05:48 -07:00
server	feat(hh): /pw command, RAM-only direnv autostart, robust lets-hack; coven→clergy	2026-05-31 22:29:17 -07:00
__init__.py	fix(security): comprehensive security hardening — TLS, HMAC WS auth, rate limiting, IP leak prevention	2026-05-25 20:30:40 -07:00