checkpoint

.gemini/agents/tier1-orchestrator.md

@@ -7,10 +7,15 @@ tools:
   - list_directory
   - discovered_tool_search_files
   - grep_search
+  - discovered_tool_get_file_summary
+  - discovered_tool_get_python_skeleton
+  - discovered_tool_get_code_outline
+  - discovered_tool_get_git_diff
   - discovered_tool_web_search
   - discovered_tool_fetch_url
   - activate_skill
   - discovered_tool_run_powershell
+  - tier3-worker
 ---
 STRICT SYSTEM DIRECTIVE: You are a Tier 1 Orchestrator.
 Focused on product alignment, high-level planning, and track initialization.

.gemini/agents/tier2-tech-lead.md

@@ -9,10 +9,16 @@ tools:
   - list_directory
   - discovered_tool_search_files
   - grep_search
+  - discovered_tool_get_file_summary
+  - discovered_tool_get_python_skeleton
+  - discovered_tool_get_code_outline
+  - discovered_tool_get_git_diff
   - discovered_tool_web_search
   - discovered_tool_fetch_url
   - activate_skill
   - discovered_tool_run_powershell
+  - tier3-worker
+  - tier4-qa
 ---
 STRICT SYSTEM DIRECTIVE: You are a Tier 2 Tech Lead.
 Focused on architectural design and track execution.

.gemini/agents/tier3-worker.md

@@ -9,6 +9,10 @@ tools:
   - list_directory
   - discovered_tool_search_files
   - grep_search
+  - discovered_tool_get_file_summary
+  - discovered_tool_get_python_skeleton
+  - discovered_tool_get_code_outline
+  - discovered_tool_get_git_diff
   - discovered_tool_web_search
   - discovered_tool_fetch_url
   - activate_skill

.gemini/agents/tier4-qa.md

@@ -7,6 +7,10 @@ tools:
   - list_directory
   - discovered_tool_search_files
   - grep_search
+  - discovered_tool_get_file_summary
+  - discovered_tool_get_python_skeleton
+  - discovered_tool_get_code_outline
+  - discovered_tool_get_git_diff
   - discovered_tool_web_search
   - discovered_tool_fetch_url
   - activate_skill

conductor/workflow.md

@@ -9,6 +9,7 @@
 5. **User Experience First:** Every decision should prioritize user experience
 6. **Non-Interactive & CI-Aware:** Prefer non-interactive commands. Use `CI=true` for watch-mode tools (tests, linters) to ensure single execution.
 7. **MMA Tiered Delegation is Mandatory:** The Conductor acts as a Tier 1/2 Orchestrator. You MUST delegate all non-trivial coding to Tier 3 Workers and all error analysis to Tier 4 QA Agents. Do NOT perform large file writes directly.
+8. **Mandatory Research-First Protocol:** Before reading the full content of any file over 50 lines, you MUST use `get_file_summary`, `get_python_skeleton`, or `get_code_outline` to map the architecture and identify specific target ranges. Use `get_git_diff` to understand recent changes.
 
 ## Task Workflow
 
@@ -22,7 +23,13 @@ All tasks follow a strict lifecycle:
 
 2. **Mark In Progress:** Before beginning work, edit `plan.md` and change the task from `[ ]` to `[~]`
 
-3. **Write Failing Tests (Red Phase):**
+3. **High-Signal Research Phase:**
+   - **Identify Dependencies:** Use `list_directory` and `grep_search` to find relevant files.
+   - **Map Architecture:** Use `get_code_outline` or `get_python_skeleton` on identified files to understand their structure.
+   - **Analyze Changes:** Use `get_git_diff` if the task involves modifying recently updated code.
+   - **Minimize Token Burn:** Only use `read_file` with `start_line`/`end_line` for specific implementation details once target areas are identified.
+
+4. **Write Failing Tests (Red Phase):**
    - **Delegate Test Creation:** Do NOT write test code directly. Spawn a Tier 3 Worker (`python scripts/mma_exec.py --role tier3-worker "[PROMPT]"`) with a prompt to create the necessary test files and unit tests based on the task criteria.
    - Take the code generated by the Worker and apply it.
    - **CRITICAL:** Run the tests and confirm that they fail as expected. This is the "Red" phase of TDD. Do not proceed until you have failing tests.

mcp_client.py

@@ -281,6 +281,31 @@ def get_code_outline(path: str) -> str:
         return f"ERROR generating outline for '{path}': {e}"
 
 
+def get_git_diff(path: str, base_rev: str = "HEAD", head_rev: str = "") -> str:
+    """
+    Returns the git diff for a file or directory.
+    base_rev: The base revision (default: HEAD)
+    head_rev: The head revision (optional)
+    """
+    import subprocess
+    p, err = _resolve_and_check(path)
+    if err:
+        return err
+    
+    cmd = ["git", "diff", base_rev]
+    if head_rev:
+        cmd.append(head_rev)
+    cmd.extend(["--", str(p)])
+    
+    try:
+        result = subprocess.run(cmd, capture_output=True, text=True, check=True, encoding="utf-8")
+        return result.stdout if result.stdout else "(no changes)"
+    except subprocess.CalledProcessError as e:
+        return f"ERROR running git diff: {e.stderr}"
+    except Exception as e:
+        return f"ERROR: {e}"
+
+
 
 # ------------------------------------------------------------------ web tools
 
@@ -407,7 +432,7 @@ def get_ui_performance() -> str:
 # ------------------------------------------------------------------ tool dispatch
 
 
-TOOL_NAMES = {"read_file", "list_directory", "search_files", "get_file_summary", "get_python_skeleton", "get_code_outline", "web_search", "fetch_url", "get_ui_performance"}
+TOOL_NAMES = {"read_file", "list_directory", "search_files", "get_file_summary", "get_python_skeleton", "get_code_outline", "get_git_diff", "web_search", "fetch_url", "get_ui_performance"}
 
 
 def dispatch(tool_name: str, tool_input: dict) -> str:
@@ -426,6 +451,12 @@ def dispatch(tool_name: str, tool_input: dict) -> str:
         return get_python_skeleton(tool_input.get("path", ""))
     if tool_name == "get_code_outline":
         return get_code_outline(tool_input.get("path", ""))
+    if tool_name == "get_git_diff":
+        return get_git_diff(
+            tool_input.get("path", ""), 
+            tool_input.get("base_rev", "HEAD"), 
+            tool_input.get("head_rev", "")
+        )
     if tool_name == "web_search":
         return web_search(tool_input.get("query", ""))
     if tool_name == "fetch_url":
@@ -551,6 +582,31 @@ MCP_TOOL_SPECS = [
             "required": ["path"],
         },
     },
+    {
+        "name": "get_git_diff",
+        "description": (
+            "Returns the git diff for a file or directory. "
+            "Use this to review changes efficiently without reading entire files."
+        ),
+        "parameters": {
+            "type": "object",
+            "properties": {
+                "path": {
+                    "type": "string",
+                    "description": "Path to the file or directory.",
+                },
+                "base_rev": {
+                    "type": "string",
+                    "description": "Base revision (e.g. 'HEAD', 'HEAD~1', or a commit hash). Defaults to 'HEAD'.",
+                },
+                "head_rev": {
+                    "type": "string",
+                    "description": "Head revision (optional).",
+                }
+            },
+            "required": ["path"],
+        },
+    },
     {
         "name": "web_search",
         "description": "Search the web using DuckDuckGo. Returns the top 5 search results with titles, URLs, and snippets. Chain this with fetch_url to read specific pages.",

mma-orchestrator/SKILL.md

@@ -26,13 +26,17 @@ If you run a test or command that fails with a significant error or large traceb
 1. **DO NOT** analyze the raw logs in your own context window.
 2. **DO** spawn a stateless Tier 4 agent to diagnose the failure.
 3. *Command:* `uv run python scripts/mma_exec.py --role tier4-qa "Analyze this failure and summarize the root cause: [LOG_DATA]"`
-4. Avoid direct reads to files, use file summaries or ast skeletons for files if they are code and we have a tool for parsing them.
+4. **Mandatory Research-First Protocol:** Avoid direct `read_file` calls for any file over 50 lines. Use `get_file_summary`, `get_python_skeleton`, or `get_code_outline` first to identify relevant sections. Use `git diff` to understand changes.
 
 ## 3. Persistent Tech Lead Memory (Tier 2)
 Unlike the stateless sub-agents (Tiers 3 & 4), the **Tier 2 Tech Lead** maintains persistent context throughout the implementation of a track. Do NOT apply "Context Amnesia" to your own session during track implementation. You are responsible for the continuity of the technical strategy.
 
-## 4. AST Skeleton Views
-To minimize context bloat for Tier 2 & 3, use "Skeleton Views" of dependencies (extracted via `mcp_client.py` or similar) instead of full file contents, unless the Tier 3 worker is explicitly modifying that specific file.
+## 4. AST Skeleton & Outline Views
+To minimize context bloat for Tier 2 & 3:
+1. Use `get_code_outline` to map out the structure of a file.
+2. Use `get_python_skeleton` to understand the interface and docstrings of dependencies.
+3. Only use `read_file` with `start_line` and `end_line` for specific implementation details once target areas are identified.
+4. Tier 3 workers MUST NOT read the full content of unrelated files.
 
 <examples>
 ### Example 1: Spawning a Tier 4 QA Agent

scripts/mma_exec.py

@@ -177,14 +177,15 @@ def execute_agent(role: str, prompt: str, docs: list[str]) -> str:
         system_directive = "STRICT SYSTEM DIRECTIVE: You are a stateless Tier 3 Worker (Contributor). " \
                            "Your goal is to implement specific code changes or tests based on the provided task. " \
                            "You have access to tools for reading and writing files (e.g., read_file, write_file, replace), " \
-                           "codebase investigation (codebase_investigator), and web tools (discovered_tool_web_search, discovered_tool_fetch_url). " \
+                           "codebase investigation (discovered_tool_get_code_outline, discovered_tool_get_python_skeleton), " \
+                           "version control (discovered_tool_get_git_diff), and web tools (discovered_tool_web_search, discovered_tool_fetch_url). " \
                            "You CAN execute PowerShell scripts via discovered_tool_run_powershell for verification and testing. " \
                            "Follow TDD and return success status or code changes. No pleasantries, no conversational filler."
     elif role in ['tier4', 'tier4-qa']:
         system_directive = "STRICT SYSTEM DIRECTIVE: You are a stateless Tier 4 QA Agent. " \
                            "Your goal is to analyze errors, summarize logs, or verify tests. " \
-                           "You have access to tools for reading files, exploring the codebase (codebase_investigator), " \
-                           "and web tools (discovered_tool_web_search, discovered_tool_fetch_url). " \
+                           "You have access to tools for reading files, exploring the codebase (discovered_tool_get_code_outline, discovered_tool_get_python_skeleton), " \
+                           "version control (discovered_tool_get_git_diff), and web tools (discovered_tool_web_search, discovered_tool_fetch_url). " \
                            "You CAN execute PowerShell scripts via discovered_tool_run_powershell for diagnostics. " \
                            "ONLY output the requested analysis. No pleasantries."
     else:
@@ -208,8 +209,8 @@ def execute_agent(role: str, prompt: str, docs: list[str]) -> str:
     # Use subprocess with input to pipe the prompt via stdin, avoiding WinError 206.
     # We use -p 'mma_task' to ensure non-interactive (headless) mode and valid parsing.
     # Whitelist tools to ensure they are available to the model in headless mode.
-    # Using 'discovered_tool_run_powershell' as it's the confirmed name for shell access.
-    allowed_tools = "read_file,write_file,replace,list_directory,glob,grep_search,discovered_tool_search_files,discovered_tool_get_file_summary,discovered_tool_run_powershell,activate_skill,codebase_investigator,discovered_tool_web_search,discovered_tool_fetch_url"
+    # Using 'discovered_tool_*' names as they are provided by tool_discovery.py
+    allowed_tools = "read_file,write_file,replace,list_directory,glob,grep_search,discovered_tool_search_files,discovered_tool_get_file_summary,discovered_tool_get_python_skeleton,discovered_tool_get_code_outline,discovered_tool_get_git_diff,discovered_tool_run_powershell,activate_skill,codebase_investigator,discovered_tool_web_search,discovered_tool_fetch_url"
     ps_command = (
         f"if (Test-Path 'C:\\projects\\misc\\setup_gemini.ps1') {{ . 'C:\\projects\\misc\\setup_gemini.ps1' }}; "
         f"gemini -p 'mma_task' --allowed-tools {allowed_tools} --output-format json --model {model}"