Openclaw Skills turns real-world social and consumer shifts into evidence-backed U.S. equity research with a skeptical, Chris Camillo-inspired workflow.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install social-arbitrage
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install social-arbitrage using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Social Arbitrage?
Openclaw Skills provides a Chris Camillo-inspired social arbitrage framework for finding underrecognized changes in consumer behavior, technology adoption, culture, weather, supply, labor, and perception, then mapping those changes to U.S.-listed common stocks and ADRs. It is designed for current, public-information research only, with a bias toward primary sources, dated evidence, and explicit falsification.
The skill does not produce trade instructions. Instead, it generates long, short, or neutral research bias for watchlists and reports after testing baseline, authenticity, independent corroboration, issuer exposure, materiality, awareness, and timing. If the evidence is weak, ambiguous, or already consensus, Openclaw Skills rejects the idea rather than forcing a ticker.
Social Arbitrage Use Cases
- Broad current trend scans across consumer, retail, apparel, beauty, food, restaurants, software, AI, semiconductors, hardware, devices, media, gaming, advertising, entertainment, healthcare, wellness, travel, mobility, housing, home improvement, finance, industrials, logistics, manufacturing, energy, commodities, weather, agriculture, labor, education, regulation, policy, and culture.
- Trend-to-ticker mapping for U.S.-listed common stocks and ADRs on NYSE, Nasdaq, or NYSE American.
- Long/short/neutral watchlists that separate meaningful social signals from noise without forcing a ticker count.
- Alternative-data research that combines social, search, news, filing, issuer, app, job, and public-agency evidence.
- Testing whether a viral post, local observation, or channel shift is material, authentic, and underrecognized.
- Clean-room public-information studies when the user may be hinting at MNPI or when the original premise needs validation.
- Producing report-ready evidence chains with as-of timestamps, direct source links, and explicit counterarguments.
How Social Arbitrage Works
- Frame the research question and set a safe default when the request is broad; do not accept the user's premise as fact, and define at least one competing hypothesis.
- Load
methodology.md,research-playbook.md, andreport-template.md; inventory available web, browser, search, news, filing, issuer, and market-data tools, and record gaps in an access ledger. - Discover observable changes outside financial media first when practical, scanning the eight default sector lenses and looking for acceleration, deceleration, spread, persistence, scarcity, discounting, waitlists, job demand, cost changes, regulation, or perception shifts.
- Establish a baseline and timestamp for every surviving observation, including the collection window, comparison basis, and what the baseline can and cannot prove.
- Test authenticity and alternatives by checking sponsorship, affiliate incentives, duplicated scripts, coordinated timing, bots, recycled content, algorithmic amplification, stock promotion, and local-event bias.
- Require at least two meaningfully independent evidence channels for A/B candidates; treat reposts, copied articles, coordinated creators, and one-dataset dashboards as non-independent.
- Map the full causal chain from observed change to behavior, then to units, price, mix, cost, retention, capacity, or perception, and onward to the exposed segment and possible reported or expectations effect.
- Verify ticker, issuer, exchange, security type, exposure path, and materiality from primary or authoritative sources before printing an A/B ticker.
- Test the information gap by checking filings, investor relations, earnings commentary, trade press, general media, financial media, analyst or estimate context when available, and recent price reaction.
- Define timing and falsification: why the evidence matters now, the likely dissemination or catalyst path, the strongest counterargument, the first inexpensive rejection test, observable falsifiers, and the next exact data or event to monitor.
- Apply all ten gates from
methodology.mdand rank without averaging away failure; evidence that contradicts any gate forces rejection, unresolved evidence lowers the tier, and A/B candidates must survive every gate provisionally. - Produce and audit the report using
report-template.md; move failed ideas to Reject, lower unresolved ideas to B or C, and use the no-qualifying-signals format when no A/B idea survives.
Social Arbitrage Setup
Openclaw Skills does not require a package install, but it does require the three reference files and tool access to current public data.
export SKILL_DIR=~/skills/social-arbitrage
mkdir -p ${SKILL_DIR}/references
test -f ${SKILL_DIR}/references/methodology.md
test -f ${SKILL_DIR}/references/research-playbook.md
test -f ${SKILL_DIR}/references/report-template.md
# Ensure the runtime has browser, search, filing, issuer, and market-data tools
# For account-gated sources, use an authorized signed-in browser session
Operationally, keep the reference files in {baseDir}/references/, use public or user-authorized sources only, and confirm the agent can browse current claims before running a research session.
Social Arbitrage Data Schema & Taxonomy
Openclaw Skills organizes research into layered artifacts rather than a single flat note.
| Artifact | Purpose | Key fields |
|---|---|---|
references/methodology.md |
Defines the ten gates and the tiering logic | gate names, A/B/C/Reject rules, rejection thresholds, no-score policy |
references/research-playbook.md |
Defines source discipline and access handling | tool inventory, access ledger rules, public-information boundaries, MNPI handling |
references/report-template.md |
Defines the final report structure | summary, evidence, mechanism, watchlist, falsifiers, links, bias labels |
| Access ledger | Records what was and was not accessible during the run | tool name, auth state, limitations, timestamp |
| Observation record | Captures each real-world signal before it is mapped to a security | claim, sector lens, date window, baseline, delta, geography, demographic context |
| Evidence bundle | Stores the support or rejection material for an observation | source URLs, channel type, independence, authenticity checks, notes, dates |
| Ticker mapping record | Connects an observation to an issuer and security | legal issuer, ticker, exchange, security type, exposure path, materiality, market cap, liquidity |
| Research verdict | Final classification for each candidate | tier, long/short/neutral research bias, counterargument, falsifiers, next monitor |
Metadata taxonomy used throughout the workflow:
- Time: as-of time, collection window, event date, publication date.
- Source type: social, search, news, filing, issuer, app, job, public-agency, industry, market-data.
- Authenticity: authentic, unresolved, or manipulated.
- Awareness: undiscovered, emerging, partially disseminated, or consensus.
- Security eligibility: common stock or ADR only; exclude OTC, funds, shells, pre-merger SPACs, illiquid microcaps, and binary pre-commercial biotechnology by default.
Generated outputs typically include a current trend report, a trend-to-ticker scan, an A/B/C/Reject watchlist, or a no-qualifying-signals report, each with direct links and dates attached to the claims they support.
Social Arbitrage Advanced Features
- Eight-lens broad scan across consumer, retail, software, AI, semiconductors, media, gaming, healthcare, travel, mobility, industrials, logistics, weather, agriculture, labor, policy, and culture.
- Access-led research discipline that records unavailable or limited channels instead of pretending a platform, post, filing, or market field was inspected.
- Independent-corroboration requirement that filters out reposts, copied articles, coordinated creators, and single-dataset dashboards.
- Baseline-first methodology that rejects static lifetime counts and caps anecdotal observations when no defensible comparison exists.
- A/B/C/Reject tiering with explicit long, short, or neutral research bias and no forced ticker count.
- Public-only, MNPI-safe clean-room mode that redirects sensitive tips into non-tradable public research.
- Exposure and materiality checks grounded in authoritative issuer, listing, segment, and liquidity evidence.
- Falsification-first workflow that surfaces counterarguments, inexpensive rejection tests, and observable falsifiers before any report is finalized.
- Built for Openclaw Skills workflows that need repeatable, evidence-backed equity research instead of hot takes or execution guidance.
- No execution, sizing, leverage, brokerage, or options instructions, keeping the output focused on research triage.
SKILL.md
Loading
Related Openclaw Skills
World Builder
A meta-skill that turns a world idea or character set into a coherent, 12-dimensional world specification.
CertainLogic Stack
CertainLogic Stack is an Openclaw Skills AI stack that combines verified knowledge, low-cost routing, caching, and audit trails to cut spend and stop hallucinations.
Aesthetic Builder
Aesthetic Builder converts a concept or character set into a structured 9D aesthetic system with dual-mode generation and extraction.
Narrative Builder
Narrative Builder turns world and character inputs into a coherent, validated story pipeline with causal logic, pacing, and theme continuity.
redact-pii
Openclaw Skills for redact-pii masks sensitive text in-line with ZeroGPU, replacing detected PII with bracketed labels before you share or log it.
Classify Structured
Schema-driven multi-axis text classification that turns arbitrary text into structured JSON labels with Openclaw Skills.
Featured*
