ed/manual_slop
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

conductor(track): Expand scope of architecture track to fully integrate MCP tools

Browse Source
This commit is contained in:
Edward R. Gonzalez
2026-03-02 12:39:41 -05:00
parent 6837a28b61
commit 9eabebf9f4
3 changed files with 17 additions and 22 deletions
Download Patch File Download Diff File Expand all files Collapse all files

25
conductor/tracks/architecture_boundary_hardening_20260302/spec.md
View File

@@ -1,30 +1,23 @@
# Track Specification: Architecture Boundary Hardening
## Overview
The `manual_slop` project serves dual roles: it is an end-user GUI application built around Human-In-The-Loop (HITL) AI orchestration, and it is the sandbox for the AI meta-tooling (`mma_exec.py`, `tool_call.py`) being used to develop it.
Because `mcp_client.py` is shared between both environments to provide robust code investigation tools, a critical HITL bypass has emerged. Additionally, the meta-tooling scripts are bleeding tokens.
The `manual_slop` project sandbox provides AI meta-tooling (`mma_exec.py`, `tool_call.py`) to orchestrate its own development. When AI agents added advanced AST tools (like `set_file_slice`) to `mcp_client.py` for meta-tooling, they failed to fully integrate them into the application's GUI, config, or HITL (Human-In-The-Loop) safety models. Additionally, meta-tooling scripts are bleeding tokens, and the internal application's state machine can deadlock.
## Current State Audit
1. **HITL Bypass in `manual_slop` Application**:
- Location: `ai_client.py` inside `_send_gemini`, `_send_gemini_cli`, `_send_anthropic`, and `_send_deepseek`.
- Issue: The `pre_tool_callback` is explicitly only checked if `name == TOOL_NAME` (which is `run_powershell`).
- If an AI agent running inside the GUI calls `set_file_slice` or `py_update_definition`, the code falls through to `elif name in mcp_client.TOOL_NAMES:` and dispatches it immediately, silently mutating the user's codebase without approval.
- *Requirement*: The application strictly requires step-by-step deterministic user approval for *any* filesystem modification, whether by script or direct AST manipulation.
1. **Incomplete MCP Tool Integration & HITL Bypass (`ai_client.py`, `gui_2.py`)**:
- Issue: New tools in `mcp_client.py` (e.g., `set_file_slice`, `py_update_definition`) are not exposed in the GUI or `manual_slop.toml` config `[agent.tools]`. If they were enabled, `ai_client.py` would execute them instantly without checking `pre_tool_callback`, bypassing GUI approval.
- *Requirement*: Expose all `mcp_client.py` tools as toggles in the GUI/Config. Ensure any mutating tool triggers a GUI approval modal before execution.
2. **Token Firewall Leak in Meta-Tooling (`mma_exec.py`)**:
- Location: `scripts/mma_exec.py:101`.
- Issue: `UNFETTERED_MODULES` hardcodes `['mcp_client', 'project_manager', 'events', 'aggregate']`. If a worker targets a file that imports `mcp_client`, the script injects the full `mcp_client.py` (~450 lines) into the context instead of its skeleton, blowing out the token budget and destroying Context Amnesia.
- Issue: `UNFETTERED_MODULES` hardcodes `['mcp_client', 'project_manager', 'events', 'aggregate']`. If a worker targets a file that imports `mcp_client`, the script injects the full `mcp_client.py` (~450 lines) into the context instead of its skeleton, blowing out the token budget.
3. **DAG Engine Blocking Stalls (`dag_engine.py`)**:
- Location: `dag_engine.py` -> `get_ready_tasks()`
- Issue: `get_ready_tasks` requires all dependencies to be explicitly `completed`. If a task is marked `blocked` (e.g. after max retries in the ConductorEngine), its dependents stay `todo` forever. The `ConductorEngine.run()` loop has no logic to handle this cleanly, causing an infinite stall.
- Issue: `get_ready_tasks` requires all dependencies to be explicitly `completed`. If a task is marked `blocked`, its dependents stay `todo` forever, causing an infinite stall.
## Desired State
- Any mutating tool from `mcp_client.py` (`set_file_slice`, `py_update_definition`, `py_set_signature`, `py_set_var_declaration`, `write_file`) must trigger a user approval dialogue, just like `run_powershell`.
- The `UNFETTERED_MODULES` list must be completely removed from `mma_exec.py` so all dependencies are reliably skeletonized.
- The `dag_engine.py` must cascade `blocked` status to downstream tasks so the track halts cleanly instead of deadlocking.
## Technical Constraints
- The UI modal must be updated or a new `pre_mutation_callback` must be introduced to handle showing the proposed AST edit vs the proposed script.
- Keep the boundary clear: changes in `ai_client.py` affect the user's `manual_slop` application experience. Changes in `mma_exec.py` affect *our* meta-tooling environment.
- All tools in `mcp_client.py` are configurable in `manual_slop.toml` and `gui_2.py`. Mutating tools must route through the GUI approval callback.
- The `UNFETTERED_MODULES` list must be completely removed from `mma_exec.py`.
- The `dag_engine.py` must cascade `blocked` status to downstream tasks so the track halts cleanly.