hermes-agent/plugins
k4z4n0v4 c1c59e3474 fix(honcho): make honcho_search do real cross-session message search
honcho_search routed through search_context() -> peer.context(search_query=),
which returns the peer's standing representation + card. The search_query arg
does not turn that endpoint into a search, so results were effectively
query-independent: the same representation blob regardless of the query. Factual
lookups ('what medication', 'which value did we pick') returned noise.

Rewire search_context() to call the workspace message-search endpoint
(Honcho.search) with a peer_perspective filter: RRF-ranked (hybrid semantic +
full-text) raw message excerpts spanning every session the peer was a member of,
across all authors, membership-time-scoped. This is the cross-session factual
recall primitive.

peer_perspective is chosen over the alternatives because it is the only scope
that is simultaneously (a) cross-session, (b) inclusive of assistant-authored
facts about the peer (peer-author search drops these, and they are a large part
of what you want to recall about yourself), and (c) privacy-scoped to the peer's
own sessions (plain workspace search leaks other peers' sessions).

- snippets are labeled by author so the model can tell user-stated facts from
  assistant-derived ones
- max_tokens is now an enforced budget (was accepted but meaningless)
- graceful fallback to peer-authored search if peer_perspective is unsupported
- query length clamped under the embedding input cap

Replaces 3 change-detector tests that asserted the old representation-dump
behavior with 4 that assert the message-search contract + fallback path.
2026-07-16 12:48:48 -07:00
..
browser fix: guard int(os.getenv()) casts against malformed env vars (#40598) 2026-06-07 06:14:24 -07:00
context_engine feat(context-engine): host contract for external context engines 2026-05-28 01:45:30 -07:00
cron_providers fix(cron/chronos): cache PyJWKClient across fires to stop JWKS fetch storm (#64641) 2026-07-15 09:27:35 +10:00
dashboard_auth feat(dashboard_auth): support confidential clients (client_secret) in self-hosted OIDC (#55344) 2026-06-30 13:32:51 +10:00
disk-cleanup docs(disk-cleanup): clarify cron output-root protection is exact-match 2026-07-01 15:42:04 +05:30
google_meet refactor(setup): route dependency installs through the canonical uv→pip→ensurepip ladder 2026-07-07 04:09:35 -07:00
hermes-achievements fix(achievements): stop card hover click loop 2026-07-16 04:34:57 -07:00
image_gen fix(deepinfra): harden multimodal provider routing 2026-07-14 02:59:39 +05:30
kanban refactor(kanban): unify attachment size cap on KANBAN_ATTACHMENT_MAX_BYTES 2026-07-16 07:33:14 -07:00
memory fix(honcho): make honcho_search do real cross-session message search 2026-07-16 12:48:48 -07:00
model-providers fix(copilot): clamp reasoning effort to the nearest supported level, not xhigh->high 2026-07-16 08:47:10 -07:00
observability fix(nemo-relay): align dynamic plugin configuration 2026-07-13 17:01:08 -06:00
platforms feat(gateway): inline choice pickers for /reasoning and /fast (Telegram, Discord, Matrix) (#65799) 2026-07-16 10:38:31 -07:00
security-guidance plugins: add security-guidance — pattern-matched warnings on dangerous code writes (#33131) 2026-05-27 02:07:21 -07:00
spotify chore: prune unused imports and duplicate import redefinitions 2026-05-28 22:26:25 -07:00
teams_pipeline fix(teams-pipeline): reject dot-only recording display_name 2026-07-01 02:03:48 -07:00
video_gen fix(video-gen): omit duration for range-based FAL families when unspecified 2026-07-16 04:24:10 -07:00
web fix(web): widen config-aware env resolution to exa/parallel/tavily/brave-free providers 2026-07-06 02:42:24 -07:00
__init__.py feat(memory): pluggable memory provider interface with profile isolation, review fixes, and honcho CLI restoration (#4623) 2026-04-02 15:33:51 -07:00
plugin_utils.py fix(plugins): thread-safe lazy-singleton helpers; fix honcho TOCTOU (#24759) (#42150) 2026-06-08 09:35:22 -07:00