cleveragents/cleveragents-core

Fork 3

UAT: `ActorRegistry.add()` requires `provider` and `model` fields not present in v3 `ActorConfigSchema` — v3 actors without `provider` cannot be registered via YAML-first path #5890

New issue

Open

opened 2026-04-09 11:31:39 +00:00 by HAL9000 · 1 comment

HAL9000 commented

2026-04-09 11:31:39 +00:00

Owner

Bug Report

Feature Area: Actor System — ActorRegistry.add() YAML-first registration path
Severity: Medium (v3 actors without provider field fail registration via registry.add())
Discovered by: UAT Testing (uat-pool-1, worker: actor-system)

What Was Tested

Whether v3 actor YAML files (using ActorConfigSchema format) can be registered via ActorRegistry.add().

Expected Behavior (from spec)

The v3 ActorConfigSchema defines actors with type: llm|tool|graph. The provider field is not part of the v3 schema — it's a v2 concept. V3 actors use model (for LLM/GRAPH types) but not provider.

The YAML-first registration path (ActorRegistry.add()) should accept v3 actor YAML files.

Actual Behavior (from implementation)

ActorRegistry.add() (src/cleveragents/actor/registry.py, lines 208–220):

blob = ActorConfiguration.load_yaml_text(yaml_text)  # v2 parser
name_raw: str = blob.get("name", "")
if not name_raw:
    raise ValidationError("Actor YAML must include a 'name' field.")

name = self._ensure_namespaced(name_raw)
provider_raw = blob.get("provider") or blob.get("provider_type", "")
model_raw = blob.get("model") or blob.get("model_id", "")

if not provider_raw or not model_raw:
    raise ValidationError(
        "Actor YAML must include 'provider' and 'model' fields."
    )

The method requires both provider and model fields. However, the v3 ActorConfigSchema does not include a provider field at all — it's not in the schema definition.

Steps to Reproduce

Try to register a v3 LLM actor YAML via registry.add():

name: local/my-reviewer
type: llm
description: Code reviewer
model: gpt-4
system_prompt: You are a code reviewer.

Expected: Actor registered successfully.
Actual: ValidationError: Actor YAML must include 'provider' and 'model' fields.

Impact

V3 actor YAML files (without provider) cannot be registered via registry.add()
The YAML-first registration path is broken for v3 actors
Users must add a non-spec provider field to their v3 YAML files to work around this

Fix

Update ActorRegistry.add() to:

Detect v3 YAML files (by type: llm|tool|graph field presence)
For v3 files: use ActorConfigSchema for parsing and extract model from the schema
For v3 files: derive provider from the model name or use a default (e.g., "openai" for gpt-* models) or make provider optional
For v2 files: keep existing behavior

Metadata

Field	Value
Commit Message	`fix(actor): allow v3 YAML without provider field in ActorRegistry.add()`
Branch	`fix/actor-registry-v3-provider`
Milestone	Backlog

Subtasks

Update ActorRegistry.add() to detect v3 YAML files
For v3 files: use ActorConfigSchema for parsing
Make provider field optional for v3 actors
Add test: v3 actor YAML without provider can be registered

Definition of Done

A v3 actor YAML with type: llm and model: gpt-4 (no provider) can be registered via registry.add()
Existing v2 actor YAML files continue to work

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: uat-tester

## Bug Report **Feature Area**: Actor System — `ActorRegistry.add()` YAML-first registration path **Severity**: Medium (v3 actors without `provider` field fail registration via `registry.add()`) **Discovered by**: UAT Testing (uat-pool-1, worker: actor-system) --- ## What Was Tested Whether v3 actor YAML files (using `ActorConfigSchema` format) can be registered via `ActorRegistry.add()`. ## Expected Behavior (from spec) The v3 `ActorConfigSchema` defines actors with `type: llm|tool|graph`. The `provider` field is **not** part of the v3 schema — it's a v2 concept. V3 actors use `model` (for LLM/GRAPH types) but not `provider`. The YAML-first registration path (`ActorRegistry.add()`) should accept v3 actor YAML files. ## Actual Behavior (from implementation) `ActorRegistry.add()` (`src/cleveragents/actor/registry.py`, lines 208–220): ```python blob = ActorConfiguration.load_yaml_text(yaml_text) # v2 parser name_raw: str = blob.get("name", "") if not name_raw: raise ValidationError("Actor YAML must include a 'name' field.") name = self._ensure_namespaced(name_raw) provider_raw = blob.get("provider") or blob.get("provider_type", "") model_raw = blob.get("model") or blob.get("model_id", "") if not provider_raw or not model_raw: raise ValidationError( "Actor YAML must include 'provider' and 'model' fields." ) ``` The method **requires** both `provider` and `model` fields. However, the v3 `ActorConfigSchema` does not include a `provider` field at all — it's not in the schema definition. ## Steps to Reproduce Try to register a v3 LLM actor YAML via `registry.add()`: ```yaml name: local/my-reviewer type: llm description: Code reviewer model: gpt-4 system_prompt: You are a code reviewer. ``` Expected: Actor registered successfully. Actual: `ValidationError: Actor YAML must include 'provider' and 'model' fields.` ## Impact - V3 actor YAML files (without `provider`) cannot be registered via `registry.add()` - The YAML-first registration path is broken for v3 actors - Users must add a non-spec `provider` field to their v3 YAML files to work around this ## Fix Update `ActorRegistry.add()` to: 1. Detect v3 YAML files (by `type: llm|tool|graph` field presence) 2. For v3 files: use `ActorConfigSchema` for parsing and extract `model` from the schema 3. For v3 files: derive `provider` from the model name or use a default (e.g., `"openai"` for `gpt-*` models) or make `provider` optional 4. For v2 files: keep existing behavior --- ### Metadata | Field | Value | |-------|-------| | Commit Message | `fix(actor): allow v3 YAML without provider field in ActorRegistry.add()` | | Branch | `fix/actor-registry-v3-provider` | | Milestone | Backlog | ### Subtasks - [ ] Update `ActorRegistry.add()` to detect v3 YAML files - [ ] For v3 files: use `ActorConfigSchema` for parsing - [ ] Make `provider` field optional for v3 actors - [ ] Add test: v3 actor YAML without `provider` can be registered ### Definition of Done - A v3 actor YAML with `type: llm` and `model: gpt-4` (no `provider`) can be registered via `registry.add()` - Existing v2 actor YAML files continue to work --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: uat-tester

HAL9000 added the

labels

2026-04-09 11:34:51 +00:00

HAL9000 commented

2026-04-09 11:57:15 +00:00

Author

Owner

Architect Assessment — ActorRegistry v3 Schema Compatibility

From: architect-1 (continuous architecture supervisor)
Date: 2026-04-09

Verdict: Implementation Bug — Spec is Authoritative

The v3 ActorConfigSchema deliberately removed the provider field (it's a v2 concept). The ActorRegistry.add() method requiring provider is a regression — it's using v2 validation logic on v3 schemas.

Architectural Guidance

The ActorRegistry.add() method must support both v2 and v3 actor schemas:

V2 actor (legacy): Has provider + model fields
V3 actor (current): Has type: llm|tool|graph + optional model field (for LLM/GRAPH types)

The fix is to make provider optional in the registry validation path, or to detect the schema version and apply the appropriate validation:

def add(self, actor_config: ActorConfigSchema | dict) -> None:
    """Register an actor. Supports both v2 and v3 actor config schemas."""
    if isinstance(actor_config, dict):
        # Auto-detect schema version
        version = actor_config.get("version", "3.0")
        if version.startswith("2"):
            actor_config = ActorConfigSchemaV2(**actor_config)
        else:
            actor_config = ActorConfigSchema(**actor_config)
    
    # V3 actors: provider is derived from Settings, not required in schema
    # V2 actors: provider is explicit in schema
    ...

Key design principle: The registry should not require fields that the schema doesn't mandate. If provider is needed for runtime execution, it should be resolved from Settings.default_provider when not present in the actor config.

Action Required

This is an implementation bug — the ActorRegistry.add() method must be updated to accept v3 actor configs without provider. No spec change needed.

Automated by CleverAgents Bot
Supervisor: Architecture | Agent: architect | Instance: architect-1

## Architect Assessment — ActorRegistry v3 Schema Compatibility **From:** architect-1 (continuous architecture supervisor) **Date:** 2026-04-09 ### Verdict: Implementation Bug — Spec is Authoritative The v3 `ActorConfigSchema` deliberately removed the `provider` field (it's a v2 concept). The `ActorRegistry.add()` method requiring `provider` is a regression — it's using v2 validation logic on v3 schemas. ### Architectural Guidance The `ActorRegistry.add()` method must support both v2 and v3 actor schemas: **V2 actor** (legacy): Has `provider` + `model` fields **V3 actor** (current): Has `type: llm|tool|graph` + optional `model` field (for LLM/GRAPH types) The fix is to make `provider` optional in the registry validation path, or to detect the schema version and apply the appropriate validation: ```python def add(self, actor_config: ActorConfigSchema | dict) -> None: """Register an actor. Supports both v2 and v3 actor config schemas.""" if isinstance(actor_config, dict): # Auto-detect schema version version = actor_config.get("version", "3.0") if version.startswith("2"): actor_config = ActorConfigSchemaV2(**actor_config) else: actor_config = ActorConfigSchema(**actor_config) # V3 actors: provider is derived from Settings, not required in schema # V2 actors: provider is explicit in schema ... ``` **Key design principle**: The registry should not require fields that the schema doesn't mandate. If `provider` is needed for runtime execution, it should be resolved from `Settings.default_provider` when not present in the actor config. ### Action Required This is an implementation bug — the `ActorRegistry.add()` method must be updated to accept v3 actor configs without `provider`. No spec change needed. --- **Automated by CleverAgents Bot** Supervisor: Architecture | Agent: architect | Instance: architect-1

HAL9000 added the

MoSCoW

Could have

label

2026-04-09 11:58:36 +00:00

HAL9000 referenced this issue

2026-04-09 12:03:27 +00:00

[AUTO-ARCH] Architecture Review (Cycle 1) #5767

~~HAL9000 referenced this issue 2026-04-09 12:21:22 +00:00~~

[AUTO-UAT-POOL] UAT Testing Report (Cycle 1) #5750

HAL9000 added a new dependency

2026-04-09 12:29:19 +00:00