Value research
Use when gathering current company facts, filings, earnings data, industry figures, or verifying a source for owner-style company research.From its SKILL.md
npx -y skills add white-source/value-grill-skill --skill value-researchAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
2.6 KB, as published. Nobody here has run it
Value Research
Discipline for gathering verifiable public information and classifying it honestly. Research informs questions and tests assumptions; it never replaces the user's board-level judgement.
Source Hierarchy
Prefer primary sources, in this order. Use secondary reporting only as a lead, never as the sole support for a material conclusion.
- Regulator filings and annual/interim reports — annual reports, 10-K/10-Q or equivalent, prospectuses, regulatory disclosures.
- Company releases and calls — earnings releases, earnings-call transcripts, investor presentations, official guidance.
- Company website — product pages, pricing, store or dealer locators, corporate governance disclosures.
- Primary industry data — trade associations, government statistics, regulatory registries, audited industry benchmarks.
- Secondary reporting — news, analyst notes, databases. Useful to locate a primary source; insufficient standing alone.
Every claim you record must carry its source title or URL and an as-of date. Timestamp any time-sensitive fact. Never state that a page you could not access was searched or verified.
Evidence Classification
Label each piece of evidence using the five classes from references/evidence-ledger.md:
| Class | Use when |
|---|---|
Fact | Source-backed, with publication date and source location |
Management statement | A company claim not independently verified |
Inference | An operating conclusion you derived from cited facts |
Assumption | A model input that could be falsified |
Unknown | Material information with no reliable support |
Never present an inference or assumption as a fact.
No-Web Fallback
When research tools are unavailable or a source cannot be reached, do not fabricate an answer. Instead output these four sections for the topic in question:
- Evidence needed — what specific fact is missing and which evidence class it would occupy.
- Primary sources to consult — the regulator filings, reports, transcripts, or registries that would settle it.
- Suggested searches — concrete queries (filing IDs, company + report type + period, industry-statistic terms).
- Next board question — the question to bring back to the boardroom that the missing evidence would answer or refine.
Hard Rule
Do not invent data, sources, access, or verification. If you did not retrieve a figure from a source, label it as an assumption or unknown and say so plainly.