@trac3er/oh-my-god 2.0.0 → 2.0.2

package/commands/OMG:stats.md

@@ -1,24 +1,225 @@
 ---
-description: "Display session analytics, tool usage trends, file hotspots, and failure patterns."
-allowed-tools: Read, Bash(bun:*), Grep
+description: "Display session analytics, tool usage trends, file heatmaps, and failure patterns."
+allowed-tools: Read, Bash(python*:*), Grep
 argument-hint: "[weekly|files|failures|dashboard]"
 ---
 
-# /OMG:stats
+# /OMG:stats — Session Analytics
 
-Use this command to inspect analytics artifacts produced by the Bun hook stack.
+Display session analytics, tool usage trends, file heatmaps, and failure patterns.
 
-## Primary data locations
+## Usage
 
-- `.omg/state/ledger/`
-- `.omg/evidence/`
-- `.omg/state/failure-tracker.json`
+```
+/OMG:stats
+/OMG:stats weekly
+/OMG:stats files
+/OMG:stats failures
+/OMG:stats dashboard
+```
 
-## Typical checks
+## Sub-Commands
 
-- current run summary and recent evidence packs
-- failure hotspots emitted by `hooks/circuit-breaker.ts`
-- file-level activity recorded by hook ledgers
-- dashboard artifacts under `.omg/state/`
+### `/OMG:stats` (default)
 
-The command is read-oriented and should not mutate tracked state except when explicitly generating a dashboard artifact under `.omg/state/`.
+Show current session summary: duration, tool calls, success rate, cost, and test runs.
+
+```python
+from hooks.query import get_session_summary
+
+summary = get_session_summary(".")
+duration_m = summary["duration"] // 60
+duration_s = summary["duration"] % 60
+print(f"Duration:       {duration_m}m {duration_s}s")
+print(f"Tool calls:     {summary['tool_calls']}")
+print(f"Success rate:   {summary['success_rate']:.1%}")
+print(f"Files modified: {summary['files_modified']}")
+print(f"Tests run:      {summary['tests_run']}")
+print(f"Tokens used:    {summary['tokens_used']:,}")
+print(f"Cost:           ${summary['cost_usd']:.4f}")
+```
+
+### `/OMG:stats weekly`
+
+Show last 7 days trend analysis by aggregating tool stats and escalation effectiveness.
+
+```python
+from hooks.query import get_tool_stats, get_escalation_effectiveness
+
+tools = get_tool_stats(".")
+escalations = get_escalation_effectiveness(".")
+
+# Per-tool breakdown
+for tool, stats in sorted(tools.items(), key=lambda x: x[1]["count"], reverse=True):
+    print(f"  {tool}: {stats['count']} calls, {stats['success_rate']:.0%} success, {stats['avg_tokens']:.0f} avg tokens")
+
+# Escalation summary
+print(f"\nEscalations: {escalations['escalations']} total, {escalations['resolved']} resolved, {escalations['unresolved']} unresolved")
+```
+
+### `/OMG:stats files`
+
+Show file-level heatmap of the most read, written, and edited files.
+
+```python
+from hooks.query import get_file_heatmap
+
+heatmap = get_file_heatmap(".")
+
+# Sort by total interactions (reads + writes + edits)
+ranked = sorted(
+    heatmap.items(),
+    key=lambda x: x[1]["reads"] + x[1]["writes"] + x[1]["edits"],
+    reverse=True,
+)
+
+print(f"{'File':<50} {'Reads':>6} {'Writes':>7} {'Edits':>6}")
+print("-" * 73)
+for path, counts in ranked[:20]:
+    total = counts["reads"] + counts["writes"] + counts["edits"]
+    print(f"{path:<50} {counts['reads']:>6} {counts['writes']:>7} {counts['edits']:>6}")
+```
+
+### `/OMG:stats failures`
+
+Show top failure patterns and their resolution status.
+
+```python
+from hooks.query import get_failure_hotspots, get_escalation_effectiveness
+
+hotspots = get_failure_hotspots(".")
+escalations = get_escalation_effectiveness(".")
+
+for pattern, details in sorted(hotspots.items(), key=lambda x: x[1]["count"], reverse=True):
+    status = "ESCALATED" if details["escalated"] else "active"
+    print(f"  [{status}] {pattern}: {details['count']} failures")
+    for err in details["last_3_errors"]:
+        print(f"    - {err[:100]}")
+
+print(f"\nResolution: {escalations['resolved']}/{escalations['escalations']} escalations resolved")
+```
+
+### `/OMG:stats dashboard`
+
+Generate a self-contained HTML dashboard at `.omg/state/dashboard.html`.
+
+The dashboard includes:
+- Session summary cards (duration, tool calls, cost, success rate)
+- Tool usage bar chart (via CDN Chart.js)
+- File heatmap table (top 20 files by interaction count)
+- Failure pattern list with escalation status
+
+```python
+import json
+import os
+from hooks.query import (
+    get_session_summary,
+    get_tool_stats,
+    get_file_heatmap,
+    get_failure_hotspots,
+)
+
+summary = get_session_summary(".")
+tools = get_tool_stats(".")
+heatmap = get_file_heatmap(".")
+failures = get_failure_hotspots(".")
+
+dashboard_data = {
+    "summary": summary,
+    "tools": tools,
+    "heatmap": dict(sorted(
+        heatmap.items(),
+        key=lambda x: x[1]["reads"] + x[1]["writes"] + x[1]["edits"],
+        reverse=True,
+    )[:20]),
+    "failures": failures,
+}
+
+# Write self-contained HTML with embedded Chart.js
+output_path = os.path.join(".omg", "state", "dashboard.html")
+os.makedirs(os.path.dirname(output_path), exist_ok=True)
+# ... (HTML template with CDN Chart.js and embedded JSON data)
+print(f"Dashboard written to {output_path}")
+```
+
+## Feature Flag
+
+- **Flag name**: `OMG_SESSION_ANALYTICS_ENABLED`
+- **Default**: `False` (disabled)
+- **Enable**: `export OMG_SESSION_ANALYTICS_ENABLED=1`
+
+Or set in `settings.json`:
+
+```json
+{
+  "_omg": {
+    "features": {
+      "SESSION_ANALYTICS": true
+    }
+  }
+}
+```
+
+## Output Example
+
+```
+============================================================
+  OMG Session Analytics — Summary
+============================================================
+
+  Duration:       23m 47s
+  Tool calls:     142
+  Success rate:   94.4%
+  Files modified: 18
+  Tests run:      6
+  Tokens used:    287,450
+  Cost:           $1.1142
+
+  Top Tools:
+    Bash:           58 calls, 95% success, 1,240 avg tokens
+    Read:           34 calls, 100% success, 890 avg tokens
+    Edit:           22 calls, 91% success, 1,100 avg tokens
+    Write:          16 calls, 100% success, 950 avg tokens
+    Grep:           12 calls, 100% success, 320 avg tokens
+
+  Failure Hotspots:
+    [ESCALATED] hooks/budget_governor.py: 4 failures
+    [active] tests/test_integration.py: 2 failures
+
+  Escalations: 1 total, 0 resolved, 1 unresolved
+
+============================================================
+```
+
+## Safety
+
+- **Read-only**: All sub-commands only read from existing ledger/tracker files
+- **Feature-gated**: Requires `SESSION_ANALYTICS` flag enabled
+- **No mutations**: Never modifies ledger data, failure tracker, or working memory
+- **Crash-isolated**: All query operations exit 0 on failure (via query.py internals)
+- **Dashboard**: Writes only to `.omg/state/dashboard.html` (inside managed state directory)
+
+## API
+
+```python
+from hooks.query import (
+    get_session_summary,
+    get_tool_stats,
+    get_failure_hotspots,
+    get_escalation_effectiveness,
+    get_file_heatmap,
+)
+
+# Current session summary
+summary = get_session_summary(".")
+
+# Per-tool statistics
+tools = get_tool_stats(".")
+
+# Failure patterns and escalation status
+failures = get_failure_hotspots(".")
+escalations = get_escalation_effectiveness(".")
+
+# File interaction heatmap
+heatmap = get_file_heatmap(".")
+```

package/commands/OMG:teams.md

@@ -1,15 +1,14 @@
 ---
-description: OMG internal team routing (standalone Bun runtime).
-allowed-tools: Read, Grep, Glob, Bash(git:*), Bash(rg:*), Bash(find:*), Bash(cat:*), Bash(bun:*)
+description: OMG internal team routing (standalone). Standalone team routing.
+allowed-tools: Read, Grep, Glob, Bash(git:*), Bash(rg:*), Bash(find:*), Bash(cat:*), Bash(python3:*)
 argument-hint: "[codex|gemini|ccg|auto] 'problem statement'"
 ---
 
-# /OMG:teams
+# /OMG:teams — Standalone Internal Router
 
-Use OMG's internal router through the Bun CLI.
+Use OMG's internal router using internal routing.
 
 ## Input contract
-
 - target: `auto|codex|gemini|ccg`
 - problem: clear issue statement
 - context: optional constraints
@@ -17,14 +16,24 @@ Use OMG's internal router through the Bun CLI.
 - expected_outcome: optional acceptance target
 
 ## Execution
+Use internal CLI router:
+
+```bash
+OMG_CLI="${OMG_CLI_PATH:-$HOME/.claude/omg-runtime/scripts/omg.py}"
+if [ ! -f "$OMG_CLI" ] && [ -f "scripts/omg.py" ]; then OMG_CLI="scripts/omg.py"; fi
+python3 "$OMG_CLI" teams --target auto --problem "[problem]"
+```
+
+For explicit target:
 
 ```bash
-OMG_CLI="${OMG_CLI_PATH:-$HOME/.claude/omg-runtime/scripts/omg.ts}"
-if [ ! -f "$OMG_CLI" ] && [ -f "scripts/omg.ts" ]; then OMG_CLI="scripts/omg.ts"; fi
-bun "$OMG_CLI" teams --target auto --problem "[problem]"
-bun "$OMG_CLI" teams --target codex --problem "[problem]"
+python3 "$OMG_CLI" teams --target codex --problem "[problem]"
 ```
 
 ## Output schema
+`TeamDispatchResult { status, findings[], actions[], evidence{} }`
 
-`TeamDispatchResult { status, target, phases[], actions[], evidence{} }`
+## Full-power sub-agent protocol
+- For non-trivial tasks, launch multiple sub-agents in parallel (`run_in_background=true`) for independent tracks.
+- Collect all task outputs before responding (`background_output` per task id).
+- Run a `sequential-thinking` merge step to produce one dependency-ordered execution plan.

package/config/lsp_languages.yaml

@@ -3,6 +3,14 @@
 # Each entry specifies server command, discovery paths, and capabilities
 
 languages:
+  - name: python
+    extensions: [.py]
+    server_command: [pylsp]
+    server_name: pylsp
+    discovery_paths: [.venv/bin/pylsp, ~/.local/bin/pylsp, /usr/local/bin/pylsp]
+    format_on_write: true
+    diagnostics_on_edit: true
+
   - name: typescript
     extensions: [.ts]
     server_command: [typescript-language-server, --stdio]