Skip to content
SkillScript-led2,382 tokens in use

News Reaction Failure Analyzer

Step two of a COT contrarian process: a Python CLI that tests whether a crowded market failed to react to favorable news and returns CONFIRMED, NOT_CONFIRMED or INSUFFICIENT_EVIDENCE.

7,428 stars on the repository MIT Licensetradermonty/claude-trading-skills · skills/news-reaction-failure-analyzer

Skill label

measured from its files

What installing News Reaction Failure Analyzer costs

183tokens idle
2,382tokens in use
7,177tokens, up to

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Up to adds the files its instructions link. Counts are approximate, made offline with cl100k_base.

Model dependence Script-led

Script-ledMixedModel-led

Most of this skill is scripts, so it behaves much the same whichever model runs it.

This describes the skill's structure, how much of it is scripts. It does not say how well it works. Scripts: scripts/analyze_news_reaction.py, scripts/reaction_math.py.

Touches

shelldetectednetworkdetectedwrites filesdetectedcredentialsdetected

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Good
  • Frontmatter12/15
  • Structure17/25
  • Examples15/15
  • Output15/15
  • Size13/15
  • Resources6/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Good
  • Runs shell commands−5

    does this on the machine, and the frontmatter does not say so

    SKILL.md, line 99
  • Reads credentials−15

    does this on the machine, and the frontmatter does not say so

    scripts/analyze_news_reaction.py, line 177
  • Reaches the network−10

    does this on the machine, and the frontmatter does not say so

    scripts/analyze_news_reaction.py, line 320
  • Writes files−8

    does this on the machine, and the frontmatter does not say so

    scripts/analyze_news_reaction.py, line 747

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Good
  • Development activity100
  • Maintenance52
  • Community68
  • Maturity36
  • Momentum40

News Reaction Failure Analyzer takes a market flagged as crowded long or short, a list of news events that Claude curates through web search, and price history from Financial Modeling Prep. It computes each event’s return and z-score, groups overlapping events into clusters, and applies a drift-significance test to judge whether the market moved against the crowd’s favorable news.

The verdict is fail-closed: too few events, stale detector data, a neutral classification or a missing price source all return INSUFFICIENT_EVIDENCE. The output describes market behaviour for research and is not investment advice.

What You Get

  • A JSON report nrf_SYMBOL_DATE.json with verdict, confidence, aggregate statistics and per-event evidence
  • A Markdown report with the evidence table, dropped events with reasons and a proxy caveat when used
  • An actual-reaction label such as FAILED_TO_RALLY, FAILED_TO_SELL_OFF, RALLIED or SOLD_OFF
  • A handoff JSON block for the contrarian-setup-gate skill
  • A documented price source fallback chain from futures symbol to ETF proxy

Common Use Cases

  • Checking whether a crowded-long market shrugged off bullish news after a COT crowding flag
  • Running a news-failure check on a single symbol with an explicit direction
  • Reviewing a verdict’s evidence table to see which events were dropped and why
  • Reusing the same test for news-failure checks outside COT, such as post-earnings drift

What Makes It Unique

The method avoids a simple failure-ratio rule. The skill explains that under pure noise most individual events fail to respond by chance, so such a rule would confirm on random data. It requires the market to have moved significantly against the favorable news, with a null false-positive rate verified by simulation and documented limits under strong correlation. Overlapping event windows are clustered, and the CLI always exits cleanly with a stated reason when evidence is missing.

Under The Hood

The skill runs a Python command-line script. It calls the Financial Modeling Prep historical price endpoint with your API key, trying the futures symbol first and an ETF proxy when the data is restricted or empty. Claude uses web search to curate the events file, with a source hierarchy from issuers and official statistics down to wire services, and is told never to invent events or URLs. The script reads the detector report and events file from disk and writes the JSON and Markdown reports to the output folder. Dependencies are listed in a requirements file.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers