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

docs(conductor): Synchronize docs for track 'MMA Tiered Architecture Verification'

Browse Source
This commit is contained in:
Edward R. Gonzalez
2026-02-25 09:02:40 -05:00
parent fd1e5019ea
commit e60eef5df8
1 changed files with 9 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
conductor/workflow.md
View File

@@ -364,15 +364,16 @@ A task is complete when:
To emulate the 4-Tier MMA Architecture within the standard Conductor extension without requiring a custom fork, adhere to these strict workflow policies: To emulate the 4-Tier MMA Architecture within the standard Conductor extension without requiring a custom fork, adhere to these strict workflow policies:
### 1. Active Model Switching (Simulating the 4 Tiers) ### 1. Active Model Switching (Simulating the 4 Tiers)
- **Activate MMA Orchestrator Skill:** To enforce the 4-Tier token firewall explicitly, invoke `/activate_skill mma-orchestrator` (or use the `activate_skill` tool) when planning or executing new tracks. - **Activate MMA Orchestrator Skill:** To enforce the 4-Tier token firewall, the agent MUST invoke `activate_skill mma-orchestrator` at the start of any implementation phase.
- **Tiered Delegation (MMA Protocol):** - **Tiered Delegation (The Role-Based Protocol):**
- **Tier 3 Worker (Implementation):** For significant code modifications (e.g., refactoring large scripts, implementing complex classes), delegate to a stateless sub-agent via `.\scripts\run_subagent.ps1 -Prompt "Modify [FILE] to implement [SPEC]..."`. Avoid performing heavy implementation directly in the primary context. - **Tier 3 Worker (Implementation):** For significant code modifications (Coding > 50 lines), delegate to a stateless sub-agent:
- **Tier 4 QA Agent (Error Analysis):** If tests fail with massive tracebacks (200+ lines), do not paste the error into the main context. Use `.\scripts\run_subagent.ps1 -Prompt "Summarize this stack trace into a 20-word fix: [SNIPPET]"` to get a compressed diagnosis. `.\scripts\run_subagent.ps1 -Role Worker -Prompt "Modify [FILE] to implement [SPEC]..."`
- **Phase Planning & Macro Merges (Tier 1):** Use high-reasoning models (e.g., Gemini 1.5 Pro or Claude 3.5 Sonnet) when running `/conductor:setup` or when reviewing a major phase checkpoint. - **Tier 4 QA Agent (Error Analysis):** If tests fail with large traces (Errors > 100 lines), delegate to a QA agent for compression:
- **Track Delegation & Implementation (Tier 2/3):** The MMA Orchestrator skill autonomously dispatches Tier 3 (Heads-Down Coding) tasks to secondary stateless instances of Gemini CLI (via `.\scripts\run_subagent.ps1 -Prompt "..."`) rather than performing heavy coding directly in the main thread. `.\scripts\run_subagent.ps1 -Role QA -Prompt "Summarize this stack trace into a 20-word fix: [SNIPPET]"`
- **QA/Fixing (Tier 4):** If a test fails with a massive traceback, **DO NOT** paste the traceback into the main conductor thread. Instead, the MMA Orchestrator skill instructs you to spawn a fast/cheap model sub-agent (via a shell command) to compress the error trace into a 20-word fix, keeping the main context clean. - **Traceability:** Use the `-ShowContext` flag during debugging to see the role-specific system prompts and hand-offs in the terminal.
- **Logging:** All hierarchical interactions are automatically recorded in `logs/mma_delegation.log` for auditable verification.
### 2. Context Checkpoints (The Token Firewall) ### 2. Context Checkpoints (The Token Firewall)
- The **Phase Completion Verification and Checkpointing Protocol** is the project's primary defense against token bloat. - The **Phase Completion Verification and Checkpointing Protocol** is the project's primary defense against token bloat.
- When a Phase is marked complete and a checkpoint commit is created, the AI Agent must actively interpret this as a **"Context Wipe"** signal. It should summarize the outcome in its git notes and move forward treating the checkpoint as absolute truth, deliberately dropping earlier conversational history and trial-and-error logs to preserve token bandwidth for the next phase. - When a Phase is marked complete and a checkpoint commit is created, the AI Agent must actively interpret this as a **"Context Wipe"** signal. It should summarize the outcome in its git notes and move forward treating the checkpoint as absolute truth, deliberately dropping earlier conversational history.
- **MMA Phase Memory Wipe:** After completing a major Phase, use the Tier 1/2 Orchestrator's perspective to consolidate state into Git Notes and then disregard previous trial-and-error histories. - **MMA Phase Memory Wipe:** After completing a major Phase, use the Tier 1/2 Orchestrator's perspective to consolidate state into Git Notes and then disregard previous trial-and-error histories.