multi-symbol-scan

Name: Multi Symbol Scan
Author: HumInvLitterae

// Scan or compare a small set of TradingView symbols with the Rust `tv` CLI. Use when the user wants a watchlist-style pass, cross-symbol quote/OHLCV comparison, technical screen, or shortlist based on currently available CLI reads.

Run Skill in Manus

$ git log --oneline --stat

stars:0

forks:0

updated:May 6, 2026 at 09:09

File Explorer

3 files

SKILL.md

readonly

package.json

"author": "HumInvLitterae"

"repository": "HumInvLitterae/tradingview-cli"

View GitHub Repository

$ install --globalskills.sh

$ download --local

Run Skill in Manus

[HINT] Download the complete skill directory including SKILL.md and all related files

Run any Skill with one click

name	multi-symbol-scan
description	Scan or compare a small set of TradingView symbols with the Rust `tv` CLI. Use when the user wants a watchlist-style pass, cross-symbol quote/OHLCV comparison, technical screen, or shortlist based on currently available CLI reads.

Multi-Symbol Scan

Use this skill to compare several TradingView symbols through the Rust tv CLI using Desktop-free comparison first. Default to Desktop-free reads for broad comparison. Move to Desktop-backed reads only for finalists that need selected-chart OHLCV, visible studies, or screenshots. Treat tv quote --source auto as hybrid source selection and tv bars as experimental lab data.

For the current observation command sequence, use docs/observation-workflows.md. This skill should stay focused on shortlist construction, not on duplicating the full source taxonomy.

Read Selection

Need	Prefer
Broad scanner discovery	`tv scanner scan` or `tv scanner hotlist`
Known-symbol first-pass comparison	`tv compare <SYMBOL>...`
Quote-only comparison	`tv quotes <SYMBOL>...`
One-symbol finalist detail	`tv snapshot <SYMBOL>`
Finalist time-window chart evidence	`tv observe chart --duration-ms ...`
Finalist chart-feed quote	`tv quote <SYMBOL> --source chart`
Visual ambiguity	`tv screenshot --region chart --output <PATH>`

Start With Scope

Confirm the symbol list, timeframe, and screening criteria from the user request.
Run tv readiness; if needed, run tv watchlist get to inspect the current TradingView watchlist.
Inspect ready, target_selection, chart_readiness, bars_readiness, and next_action_hint before assuming the Desktop session is usable. If more than one chart target is open, run tv tab list and use target_cli_args, for example tv --target-id <ID> ..., for any chart-specific follow-up. Do not use TV_CDP_TARGET_ID.
For broad discovery, prefer tv scanner hotlist or tv scanner scan before mutating the chart across many symbols.
Keep chart-by-chart inspection small and serial. The Rust CLI does not implement the old MCP batch_run helper.

Scan Workflow

Use tv scanner hotlist <SLUG> --limit <N> or tv scanner scan ... for broad read-only discovery when the criteria can be expressed as scanner filters. Use tv scanner metainfo --field <FIELD> when you need to confirm scanner field availability.
For several known symbols, use tv compare <SYMBOL>... for an ordered Desktop-free comparison packet with quote, info, and fundamentals sections. Read summary for requested/resolved/error counts, section success counts, missing totals, requested-to-resolved symbol mapping, field coverage, and requested-order indexes. summary.coverage_status is only evidence completeness (complete, partial, or blocked); keep raw items[] as the evidence source for any actual comparison. Per-item follow_up_hints are available next evidence surfaces, not recommendations. Stable follow-up kinds are snapshot, chart_quote, observe_chart, and screenshot; keep chart_quote as the canonical selected-chart quote kind and do not rename it to quote_chart. Use items[].missing_evidence[] to route section gaps to stable follow-up kinds such as snapshot or chart_quote; do not treat those hints as symbol ranking. Use tv quotes <SYMBOL>... only when ordered quote fields are enough. Use tv snapshot <SYMBOL> for one-symbol first-pass evidence before mutating the chart; it combines quote, symbol info, and fundamentals sections. Snapshot summary and missing_evidence[] provide one-symbol coverage and follow-up routing readback, while raw sections remain the evidence. Use lower-level tv quote, tv info, or tv fundamentals <SYMBOL> --group earnings|valuation|dividends|financials only when one section is enough. Preserve source_category: "desktop_free_read", requires_desktop: false, and non_mutating: true when reporting this REST-backed evidence.
Treat scanner-backed price reads as screening evidence rather than a realtime entitlement guarantee. Use tv quote <SYMBOL> --source chart only for symbols where the selected TradingView Desktop chart feed matters. Do not implement manual sleep or double-call workarounds; chart-source quote readiness is handled by the CLI with consecutive stable samples and will fail if stale chart bars remain. Do not build multi-symbol realtime loops on chart-source quote; concurrent or nearby chart mutations can still contend with the visible chart. Prefer Desktop-free reads for broad comparison, then use chart-source quote only for finalist symbols that need the selected Desktop feed. Do not use chart-source quote for premarket or postmarket fields; use scanner-backed tv quote, tv quotes, tv snapshot, or tv compare when extended-hours evidence matters. Desktop quote-session probes can expose pre/post field names, but they are phase-sensitive live evidence and not a stable multi-symbol screening source. A post-market phase observation does not make quote-session pre/post close fields equivalent to scanner extended_hours. Use TV_EXPERIMENTAL_BARS=1 tv bars <EXCHANGE:SYMBOL> --count <N> only as a lab-gated browserless bars check when experimental WebSocket data is acceptable; keep it separate from stable tv ohlcv chart evidence.
Set the timeframe once with tv timeframe <RESOLUTION> when the scan uses a shared timeframe.
Switch the chart with tv symbol <SYMBOL> only when OHLCV, visible studies, drawings, or screenshots are needed. After switching, confirm fresh chart data with tv ohlcv --count 1 or tv ohlcv --summary.
Gather tv ohlcv --summary, tv values, and drawing-derived reads such as tv data lines, tv data labels, or tv data boxes only for symbols that need chart context.
Use tv observe chart --duration-ms <MS> --heartbeat-ms <MS> for a bounded readiness-plus-last-bar observation window after the scan identifies symbols worth watching. Use lower-level tv stream quote, tv stream bars, or tv stream all only when a specific stream sample type is needed. Read JSONL source_category, requires_desktop, and non_mutating metadata before comparing those observations with Desktop-free scanner or quote results. Apply the same source metadata check to tv state, tv ohlcv, and chart-source quote payloads.
After user approval, add selected symbols with tv watchlist add-bulk <SYMBOL>... --allow-partial; it inherits the API-backed single-symbol add path and reports duplicates or partial failures.
Capture screenshots selectively for finalists or ambiguous cases with tv screenshot --region chart --output <PATH>. Treat screenshot output as Desktop-backed visual evidence with writes_file: true, not as a market data source.
Present a shortlist or candidate comparison and explain which observations came from scanner REST data, chart reads, or visual interpretation. Do not imply that tv compare itself ranked or recommended the symbols.

Use market-data-interpretation when quote freshness, source differences, extended-hours fields, fundamentals, earnings date/time fields, or missing values matter. Use screener-result-analysis when explaining why scanner or Screener rows matched a screen.

OHLCV Recovery

If a finalist chart read returns an ohlcv failure while symbol or quote reads still work, do not keep retrying the same target. Preserve the full JSON error envelope, inspect error.details.phase, bar_index_state, and next_action_hint, rerun tv readiness, choose the active chart target's target_cli_args, run tv --target-id <ID> state, and retry tv --target-id <ID> ohlcv --count 1. If structured target and chart readiness fields do not explain the visible state, capture tv screenshot --region chart --output <PATH> or ask the user to inspect the chart.

If the current environment is the Codex app and Computer Use is available, it can be used as an optional visual inspection aid after the CLI checks. Do not assume it is available for Codex CLI, packaged agents, or other CLI-only runtimes.

Use tv timeframe <RESOLUTION> for shared timeframe setup. tv interval is not a command. Use tv compare <SYMBOL>... for multi-symbol Desktop-free evidence, tv snapshot <SYMBOL> for one-symbol Desktop-free quote, info, and fundamentals context, including coverage and missing-evidence readback, and tv info <SYMBOL> or tv fundamentals <SYMBOL> with --group or --field only when that narrower section is all you need. Use tv info without a symbol only for current-chart metadata.

Boundaries

The Rust CLI can inspect and mutate the current watchlist, read scanner REST data, read chart-model data, and stream read-only chart samples as JSONL. Watchlist add/remove prefer API-backed mutation with readback checks, but still require user approval because they change account state. The CLI does not compute arbitrary historical indicator series, run strategy batches, or provide a generic batch-run helper.

Read references/workflow.md when the task needs the original MCP scan shape translated into the current CLI surface.

name	multi-symbol-scan
description	Scan or compare a small set of TradingView symbols with the Rust `tv` CLI. Use when the user wants a watchlist-style pass, cross-symbol quote/OHLCV comparison, technical screen, or shortlist based on currently available CLI reads.