MCP bridge for TradingView Desktop via Chrome DevTools Protocol
43 tools with significant definition gaps. Most tools have basic descriptions (10-50 chars) but lack parameter schemas, type definitions, and output documentation. No tool annotations (readOnlyHint/destructiveHint). Parameter descriptions are minimal or absent. Error handling and recovery guidance missing. Naming is verb-noun consistent but descriptions do not explain WHEN to use each tool or dependencies between them. No pagination documented despite tools returning potentially large datasets (e.g., data_get_ohlcv, data_get_pine_tables). Output schemas not documented anywhere in source.
create alert
delete alert
list alerts
run action across multiple symbols/timeframes
take screenshot of chart regions
get symbol, timeframe, all indicator names + entity IDs
add/remove studies. USE FULL NAMES: "Relative Strength Index" not "RSI"
jump to a date (ISO format)
change ticker
No input schemas visible in source code. 43 tools lack JSON Schema definitions with type and description for parameters. Cannot verify parameter constraints, enums, or validation rules.
Output schemas not documented. No tool returns a documented schema showing what fields the LLM should expect. Agents cannot plan downstream calls or extract required IDs for chaining.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 41 | <=2025-11-25 | v2 |
change resolution
change chart style
get price bars
price zones as {high, low} pairs
text annotations with prices ("PDH 24550", "Bias Long", etc.)
horizontal price levels from custom indicators (deduplicated, sorted)
table data as formatted rows (session stats, analytics dashboards)
get current numeric values from ALL visible indicators (RSI, MACD, BB, EMA, etc.)
draw shapes on chart
check server health and CDP connection status
change indicator settings (length, source, etc.)
focus pane
list panes
set pane layout
set symbol for pane
read log output
read errors
read Pine Script source (WARNING: can return 200KB+ for complex scripts)
inject code
compile + check errors
get real-time price snapshot (last, OHLC, volume)
start replay mode
get replay status
step through replay
stop replay mode
execute trade in replay
close tab
list tabs
create new tab
switch to tab
auto-detect and start TradingView with CDP on any platform
add symbol to watchlist
get watchlist data
remove symbol from watchlist
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). Risk metadata is provided in the tool list but not exposed in MCP tool definitions. LLMs cannot determine which tools are safe to retry or have side effects.
Descriptions lack WHEN/WHY guidance. Most descriptions state WHAT the tool does (e.g., 'get symbol, timeframe') but not WHEN to call it vs similar tools or what prerequisites exist. LLMs cannot disambiguate between data_get_pine_lines, data_get_pine_labels, data_get_pine_tables, data_get_pine_boxes.
No error handling or recovery guidance. Tools do not document what errors can occur, how to interpret them, or what the LLM should do next. E.g., chart_manage_indicator says 'add/remove studies' but does not explain what happens if the indicator name is invalid or already exists.