US Stock Market Sentiment Monitor
Judges whether the US market is in greed or fear using 5 core indicators, then delivers actionable position recommendations.
Pure prompt-framework skill with no file writes, command execution, or external side effects; relies only on web_search for public market data, with transparent data flow, source links in reports, disclaimers, and an explicit instruction to admit missing data rather than guess. Deductions: a mandatory promotional footer must be appended to every output — an unnecessary injection into user-facing content that weakens output neutrality and transparency; unverified publisher leaves attribution evidence limited.
The five-indicator framework is internally consistent with clear execution steps, abnormal-input handling (mark 'Data currently unavailable', cite most recent data), and a structured output template; the happy path is plausible. Deductions: entirely dependent on real-time web_search results with no tests or reproducible key paths; several warning thresholds are vague ('85th historical percentile', 'extreme levels since 2007'), so different runs may reach different conclusions; determinism and failure-feedback quality are thin.
Trigger scenarios are well described (US stock overheating, position reduction, sentiment indicators), non-fit boundaries are explicit (weekly/monthly medium-term only, not for intraday), multi-language output is supported, and disclaimers are clear. Deductions: core function fully depends on web_search access to overseas financial data sources (NAAIM, JPMorgan, Goldman research), which is uncertain from mainland-China networks; some indicators (hedge fund leverage) are not real-time public, so post-trigger completion is not guaranteed.
Documentation is well layered (framework → indicators → scoring table → output template → steps → caveats), with an MIT license, bilingual README, and complete install instructions. Deductions: this skill itself has no version number or changelog; no FAQ or troubleshooting; maintenance responsibility and update path are only implicit in individual authorship, not explicitly stated.
Provides a structured sentiment framework and directly usable report template, offering marginal value over manual searching, with honest disclosure of data lag and limits. Deductions: static review cannot verify output quality; key judgments depend on AI search and interpretation; position-reduction advice (10-30%) is coarse and outputs still need user verification and judgment.
Requires primary source links at the end of each report; indicator definitions and thresholds are publicly checkable; fact and inference are largely separated. Deductions: no tests, CI, or reproducible execution evidence; all thresholds (85th percentile, 22-23x PE) are author claims without independent corroboration; no auditable sample outputs.
- This is a static source review only; nothing was executed and output quality is unverified.
- A promotional footer is forcibly appended to every output; users should distinguish analysis content from promotion.
- All indicator thresholds are author claims without verification; position advice is reference only, not investment advice.
- Core function depends on real-time web search of overseas financial data sources, which may be unreachable from mainland-China networks; some indicators (e.g., hedge fund leverage) are not real-time public and searches may fail.
- Several warning criteria (85th percentile, historical extremes) are vaguely defined; different runs may reach different conclusions.
What does this skill do, and when should you use it?
This is one of five independent skills bundled in the Day1Global-Skills repo (us-market-sentiment), a sentiment monitoring and position recommendation system for US equity investors. It tracks 5 core indicators: the NAAIM Exposure Index, institutional equity allocation, retail net buying, S&P 500 forward P/E, and hedge fund leverage. Each indicator has explicit warning thresholds (e.g., forward P/E near 22-23x triggers a valuation warning), and the number of indicators in warning status maps to a six-level sentiment rating — from Panic (consider adding positions) to Extreme Greed with all 5 triggered (cut 20-30% or hedge). The skill fetches latest data via web_search and produces a structured report with source links. All output is for reference only and does not constitute investment advice.
At runtime it runs web_search queries for each of the 5 indicators (e.g., 'NAAIM exposure index this week', 'JPMorgan retail investor net buying'); scores each one as normal/warning/contrarian signal against thresholds built into the SKILL.md; counts triggered warnings and maps them to a six-level rating table (0 warnings → Panic, consider adding; 3 → Greed, trim 10-20%; 5 → full warning, significantly reduce or hedge); then generates a Markdown report from a fixed template containing an indicator dashboard, comprehensive rating, position recommendation, and disclaimers, appending primary data-source links. If an indicator's data cannot be found, it states 'Data currently unavailable' rather than guessing.
- A US equity holder sensing the market is overheated wants to know whether — and by how much — to reduce positions
- An individual investor wants a quick read on where institutional and retail positioning currently stands
- An investor worried about valuation bubbles wants to check whether S&P 500 forward P/E is near the 2000 dot-com or 2021 peaks
- A medium-term investor tracking hedge fund leverage crowding wants to gauge forced-liquidation risk on a pullback
- A contrarian-minded investor looking for bottom-zone signals when panic indicators cluster
What are this skill's strengths and limitations?
- Fully transparent framework: warning thresholds, the six-level rating table, and the output template are all spelled out in SKILL.md and easy to audit or customize
- Multi-dimensional indicator set covering institutions (NAAIM, State Street allocation, hedge fund leverage), retail flows, and valuation — richer than a single fear/greed index
- Explicit handling of contrarian signals: heavy institutional de-risking or retail panic selling is flagged as a potential entry opportunity
- Requires honest 'data unavailable' labeling with latest known values and forces source links, aiding verification
- Heavily dependent on a web_search tool; clients without web access cannot fetch current data and the skill effectively fails
- Some data (hedge fund leverage, institutional allocation) has no real-time public source and relies on news reports, possibly 1-2 weeks stale
- Thresholds (e.g., forward P/E 22-23x, retail buying above the 85th percentile) are the authors' rule-of-thumb values with no cited derivation
- No test suite or historical backtesting evidence is provided, so the framework's track record is unverifiable
- Repo has no topics and the docs state no maintenance cadence beyond the MIT license
How do you install this skill?
Method 1 (auto-discovery): run npx skills add https://github.com/star23/Day1Global-Skills --all; skills install to .agents/skills/ and activate automatically on matching topics. Note this installs the whole 5-skill collection — the repo documents no single-skill install command. Method 2 (Claude Code slash commands): git clone https://github.com/star23/Day1Global-Skills.git, then cp -r Day1Global-Skills/us-market-sentiment your-project/.claude/skills/ (or to ~/.claude/skills/ for global access), restart Claude Code, and invoke via /us-market-sentiment.
How do you use this skill?
No configuration needed after install — just ask in natural language, e.g., 'Is the US stock market overheating right now?', 'Should I reduce my positions?', 'What is the current market sentiment?'. The skill pulls the latest data for all 5 indicators via web_search and generates a monitoring report. Report language matches your input language. The authors note these indicators suit weekly/monthly medium-term assessment, not intraday decisions.