UAT: `agents plan guard` command not implemented — spec v3.5.0 deliverable #12 #5188

New issue

Open

opened 2026-04-09 03:02:38 +00:00 by HAL9000 · 0 comments

HAL9000 commented

2026-04-09 03:02:38 +00:00

Owner

Bug Report

Feature Area: Plan Lifecycle — agents plan guard CLI command
Severity: High — entire command is missing; spec deliverable #12 for v3.5.0 is unmet
Source: src/cleveragents/cli/commands/plan.py — command not present

What Was Tested

Code-level analysis of the plan CLI commands to verify all spec-required commands are implemented.

Expected Behavior (from spec §CLI Commands — plan guard, line 46897)

The spec defines agents plan guard as a required deliverable for v3.5.0:

Deliverable #12: agents plan guard shows active guards for a plan
Verifiable Check: Command shows denylist, budget caps, tool limits

The command should display the active AutomationGuard configuration for a plan, including:

Tool denylist: Tools that are blocked for this plan
Tool allowlist: Tools that are permitted (if configured)
Budget caps: max_total_cost and max_cost_per_plan limits
Tool call limits: max_tool_calls_per_step
Write approval: Whether write operations require approval
Apply approval: Whether apply phase requires approval

Actual Behavior

The command does not exist. Running agents plan guard would produce:

Error: No such command 'guard'.

The AutomationGuard model is fully implemented in src/cleveragents/domain/models/core/automation_guard.py and the AutomationProfile.check_guard() method is functional, but there is no CLI command to inspect the active guards for a running plan.

Code Location

src/cleveragents/cli/commands/plan.py — no @app.command("guard") exists
src/cleveragents/domain/models/core/automation_guard.py — AutomationGuard model is implemented
src/cleveragents/domain/models/core/automation_profile.py — AutomationProfile.check_guard() is implemented

Impact

v3.5.0 deliverable #12 is unmet
Users cannot inspect active guard constraints for a plan without reading the automation profile directly
Debugging guard enforcement issues requires manual inspection of the automation profile
The agents plan guard command is referenced in the spec's acceptance criteria for the Guard Enforcement epic (#4951)

Expected Command Signature

Based on the spec pattern for similar commands:

agents plan guard <PLAN_ID>

Expected output (rich format):

╭─ Active Guards ─────────────────────────────────────────╮
│ Plan: <PLAN_ID>                                          │
│ Profile: trusted                                         │
│                                                          │
│ Tool Denylist: (none)                                    │
│ Tool Allowlist: (all allowed)                            │
│ Max Tool Calls/Step: (unlimited)                         │
│ Max Total Cost: $5.00                                    │
│ Require Write Approval: no                               │
│ Require Apply Approval: yes                              │
╰──────────────────────────────────────────────────────────╯

Parent Epic

Part of #4951 (EPIC: Guard Enforcement & Safety Profiles)

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

## Bug Report **Feature Area:** Plan Lifecycle — `agents plan guard` CLI command **Severity:** High — entire command is missing; spec deliverable #12 for v3.5.0 is unmet **Source:** `src/cleveragents/cli/commands/plan.py` — command not present --- ## What Was Tested Code-level analysis of the plan CLI commands to verify all spec-required commands are implemented. ## Expected Behavior (from spec §CLI Commands — plan guard, line 46897) The spec defines `agents plan guard` as a required deliverable for v3.5.0: > **Deliverable #12**: `agents plan guard` shows active guards for a plan > **Verifiable Check**: Command shows denylist, budget caps, tool limits The command should display the active `AutomationGuard` configuration for a plan, including: - **Tool denylist**: Tools that are blocked for this plan - **Tool allowlist**: Tools that are permitted (if configured) - **Budget caps**: `max_total_cost` and `max_cost_per_plan` limits - **Tool call limits**: `max_tool_calls_per_step` - **Write approval**: Whether write operations require approval - **Apply approval**: Whether apply phase requires approval ## Actual Behavior The command does not exist. Running `agents plan guard` would produce: ``` Error: No such command 'guard'. ``` The `AutomationGuard` model is fully implemented in `src/cleveragents/domain/models/core/automation_guard.py` and the `AutomationProfile.check_guard()` method is functional, but there is no CLI command to inspect the active guards for a running plan. ## Code Location - `src/cleveragents/cli/commands/plan.py` — no `@app.command("guard")` exists - `src/cleveragents/domain/models/core/automation_guard.py` — `AutomationGuard` model is implemented - `src/cleveragents/domain/models/core/automation_profile.py` — `AutomationProfile.check_guard()` is implemented ## Impact - v3.5.0 deliverable #12 is unmet - Users cannot inspect active guard constraints for a plan without reading the automation profile directly - Debugging guard enforcement issues requires manual inspection of the automation profile - The `agents plan guard` command is referenced in the spec's acceptance criteria for the Guard Enforcement epic (#4951) ## Expected Command Signature Based on the spec pattern for similar commands: ``` agents plan guard <PLAN_ID> ``` Expected output (rich format): ``` ╭─ Active Guards ─────────────────────────────────────────╮ │ Plan: <PLAN_ID> │ │ Profile: trusted │ │ │ │ Tool Denylist: (none) │ │ Tool Allowlist: (all allowed) │ │ Max Tool Calls/Step: (unlimited) │ │ Max Total Cost: $5.00 │ │ Require Write Approval: no │ │ Require Apply Approval: yes │ ╰──────────────────────────────────────────────────────────╯ ``` ## Parent Epic Part of #4951 (EPIC: Guard Enforcement & Safety Profiles) --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: uat-tester

HAL9000 added the

labels

2026-04-09 03:02:41 +00:00

HAL9000 referenced this issue

2026-04-09 03:08:46 +00:00

[AUTO-UAT-POOL] UAT Testing Report (Cycle 4) #5160