From cc697e3771aee1a7697c2e116751b505dcb3039c Mon Sep 17 00:00:00 2001 From: Jeffrey Phillips Freeman Date: Wed, 11 Feb 2026 10:27:53 -0500 Subject: [PATCH] Docs: Updated the section describing missing sections in the document. --- docs/specification.md | 22 +++++++++++----------- 1 file changed, 11 insertions(+), 11 deletions(-) diff --git a/docs/specification.md b/docs/specification.md index f96fa2df0..fc6006c6f 100644 --- a/docs/specification.md +++ b/docs/specification.md @@ -23895,10 +23895,10 @@ CleverAgents should define where each concept lives: * Actions: stored in a registry (local files or server DB) * Actors: stored similarly (config files + DB indexing) -* Projects: stored locally or on server -* Plans: stored in plan DB with full logs -* Context indexes: vector store / graph store / SQLite -* Artifacts: filesystem or object storage +* Projects: stored locally or on server (depending on namespace) +* Plans: stored in plan DB with full logs, may be stored in the local or remote DB depending on namespace. +* Context indexes: vector store / graph store / Database (keeping in mind these use abstraction layers and should support any backend for each type of store). +* Artifacts: filesystem or object storage, depending on the namespace of the project it is attached to it may be stored locally or on the server. ### Observability @@ -23908,21 +23908,21 @@ To debug large plans: * every actor call should log prompt/context references * every tool call should log resource access (with parent skill noted) * every checkpoint should be recorded +* Extensive use of langChain wherever relevant to ensure we leverage third-party libraries to do the heavy lifting ### Security Model * sandbox isolation -* resource-level ACLs +* resource-level ACLs for remote resources on the server (based on namespace), local resources should have unrestricted access in terms of permissions, though access may be restricted when we want to limit the actors to read-only or idempotent interactions, etc. * prompt injection mitigations (server mode) -* secret management (API keys, DB credentials) -* audit logs for apply +* secret management (API keys, DB credentials), which needs to be addressed both on the server and locally. +* audit logs for apply, this should be stored either on the server or locally depending on the namespace for the plan and possibly the project(s). ### Extensibility -* plugin system for skills -* custom node types -* action templates -* actor templates (noted as missing currently) +* plugin system for skills / tools +* custom backends (for example the vector store) +* Any other points of extensibility we may not have already covered. ## Summary of Key Intended Behaviors (If You Only Read One Section)