UAT: `agents session tell` is stubbed — orchestrator actor not invoked, spec requires real LLM execution #4862

New issue

Open

opened 2026-04-08 20:10:13 +00:00 by HAL9000 · 0 comments

HAL9000 commented

2026-04-08 20:10:13 +00:00

Owner

Bug Report

Feature Area: Session Management — agents session tell
Severity: High (core feature non-functional)
Found by: UAT tester, code analysis

What Was Tested

The agents session tell command is the primary user interface for CleverAgents per the specification. It is supposed to route the user's natural-language prompt to the bound orchestrator actor, which then interprets the request and issues CleverAgents commands (creating actions, plans, project changes, etc.).

Expected Behavior (from spec)

Per docs/specification.md §agents session tell:

This is the main natural-language interface for interacting with CleverAgents. The orchestrator interprets your request and issues the necessary CleverAgents commands under the hood — creating actions, plans, or project changes as needed.

The spec shows the orchestrator executing real commands:

Commands Executed
  - agents action create --config ./actions/refresh-locks.yaml
  - agents resource add git-checkout local/platform-repo --path /repos/platform
  - agents project link-resource local/platform local/platform-repo

And returning structured output including:

plan_request block (actor, session, automation, prompt)
commands_executed list
result block (action, project, resource)
usage block (input_tokens, output_tokens, cost, duration_s, tool_calls)

Actual Behavior

The implementation in src/cleveragents/cli/commands/session.py (lines 813–853) is explicitly stubbed:

# Stub actor execution: generate simple assistant response
assistant_content = (
    f"Acknowledged: {prompt[:100]}"
    if not actor
    else f"[{actor}] Acknowledged: {prompt[:100]}"
)

The docstring even states: "For M3, the actor execution is stubbed — the assistant echoes an acknowledgement."

The command:

Appends the user message to the session ✓
Appends a hardcoded "Acknowledged: {prompt}" assistant message ✗
Does NOT invoke the bound orchestrator actor ✗
Does NOT route through A2A / SessionWorkflow.tell() ✗
Does NOT execute any CleverAgents commands ✗
Does NOT return usage metrics ✗

Code Location

src/cleveragents/cli/commands/session.py, lines 792–864 (tell function)
The A2A session/prompt operation (SessionWorkflow.tell()) is defined in the spec at §A2A but not wired up to the CLI session tell command

Steps to Reproduce

# Create a session
agents session create --actor local/orchestrator

# Try to use it as the primary interface
agents session tell --session <SESSION_ID> "Create an action to refresh dependency locks"

# Actual output: "Acknowledged: Create an action to refresh dependency locks"
# Expected output: Orchestrator executes commands and returns structured result

Impact

This is the primary user interface for CleverAgents. Without real orchestrator execution, the entire natural-language workflow is non-functional. Users cannot use session tell to create actions, plans, or project changes.

Definition of Done

session tell routes the prompt through the bound orchestrator actor via A2A message/send or SessionWorkflow.tell()
The orchestrator's response is persisted as an assistant message in the session
The command output matches the spec format (Plan Request panel, Commands Executed panel, Result panel, Usage panel)
--stream flag streams the response token-by-token
--actor override works for per-request actor switching

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

## Bug Report **Feature Area:** Session Management — `agents session tell` **Severity:** High (core feature non-functional) **Found by:** UAT tester, code analysis --- ### What Was Tested The `agents session tell` command is the **primary user interface** for CleverAgents per the specification. It is supposed to route the user's natural-language prompt to the bound orchestrator actor, which then interprets the request and issues CleverAgents commands (creating actions, plans, project changes, etc.). ### Expected Behavior (from spec) Per `docs/specification.md` §`agents session tell`: > This is the main natural-language interface for interacting with CleverAgents. The orchestrator interprets your request and issues the necessary CleverAgents commands under the hood — creating actions, plans, or project changes as needed. The spec shows the orchestrator executing real commands: ``` Commands Executed - agents action create --config ./actions/refresh-locks.yaml - agents resource add git-checkout local/platform-repo --path /repos/platform - agents project link-resource local/platform local/platform-repo ``` And returning structured output including: - `plan_request` block (actor, session, automation, prompt) - `commands_executed` list - `result` block (action, project, resource) - `usage` block (input_tokens, output_tokens, cost, duration_s, tool_calls) ### Actual Behavior The implementation in `src/cleveragents/cli/commands/session.py` (lines 813–853) is explicitly stubbed: ```python # Stub actor execution: generate simple assistant response assistant_content = ( f"Acknowledged: {prompt[:100]}" if not actor else f"[{actor}] Acknowledged: {prompt[:100]}" ) ``` The docstring even states: *"For M3, the actor execution is stubbed — the assistant echoes an acknowledgement."* The command: 1. Appends the user message to the session ✓ 2. Appends a hardcoded "Acknowledged: {prompt}" assistant message ✗ 3. Does NOT invoke the bound orchestrator actor ✗ 4. Does NOT route through A2A / `SessionWorkflow.tell()` ✗ 5. Does NOT execute any CleverAgents commands ✗ 6. Does NOT return usage metrics ✗ ### Code Location - `src/cleveragents/cli/commands/session.py`, lines 792–864 (`tell` function) - The A2A `session/prompt` operation (`SessionWorkflow.tell()`) is defined in the spec at §A2A but not wired up to the CLI `session tell` command ### Steps to Reproduce ```bash # Create a session agents session create --actor local/orchestrator # Try to use it as the primary interface agents session tell --session <SESSION_ID> "Create an action to refresh dependency locks" # Actual output: "Acknowledged: Create an action to refresh dependency locks" # Expected output: Orchestrator executes commands and returns structured result ``` ### Impact This is the **primary user interface** for CleverAgents. Without real orchestrator execution, the entire natural-language workflow is non-functional. Users cannot use `session tell` to create actions, plans, or project changes. ### Definition of Done - [ ] `session tell` routes the prompt through the bound orchestrator actor via A2A `message/send` or `SessionWorkflow.tell()` - [ ] The orchestrator's response is persisted as an assistant message in the session - [ ] The command output matches the spec format (Plan Request panel, Commands Executed panel, Result panel, Usage panel) - [ ] `--stream` flag streams the response token-by-token - [ ] `--actor` override works for per-request actor switching --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: uat-tester

HAL9000 referenced this issue

2026-04-08 20:13:13 +00:00

[Automated] CleverAgents Build Session - 2026-04-08 #4799

HAL9000 referenced this issue

2026-04-08 20:16:04 +00:00

[Automated] CleverAgents Build Session - 2026-04-08 #4799

HAL9000 added this to the v3.3.0 milestone

2026-04-08 20:17:11 +00:00

HAL9000 referenced this issue

2026-04-08 20:19:20 +00:00

[Automated] CleverAgents Build Session - 2026-04-08 #4799

HAL9000 referenced this issue