agentleFS
Sign inStart free
reutavidanreutavidan/zetesis-plugin/skills/ground/SKILL.md

ground

Skill0 starsChanged 3 months agoMITView on GitHub

Check the empirical claims in a document against retrieved literature and report which hold, which are thin, and which have no support.

7 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions7 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 7

    1. ilanbm/kpopper★ 12same name
    2. synaptiai/agent-capability-standard★ 6same name
    3. envisioning/skills★ 2same name
    4. ahelme/skills★ 1same name
    5. yiyaw-lab/agent-armor★ 1same name
    6. halakou/ground★ 0same name
    7. tamesystems/intent-engineering-skills★ 0same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository4 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

3 rules. 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 7 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If that names a file, read it.

    line 11
  2. Distinctive:

    if it fails, what the evidence actually says instead

    line 40
  3. Distinctive:

    Do not rewrite the user's document unless they ask.

    line 48

“Common” is measured: kept by at least 6 in 10 of the 7 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Ground a document's claims

Target: $ARGUMENTS

If that names a file, read it. If it is a claim written out, work on that. If it is empty, use the document currently in context, and ask which one only if that is genuinely ambiguous.

What to do

  1. Extract the empirical claims. Statements that assert something checkable about the world: an effect, a magnitude, a mechanism, a comparison, a rate. Skip definitions, opinions, and statements of intent. Aim for the load-bearing ones rather than every sentence; a claim is load-bearing if the argument fails when it fails. List them for the user before you start checking, numbered.
  2. Work out the era. For each claim, decide the year it should be judged against. Usually that is the publication year of the source it rests on, or the year the document is about. Pass it as as_of. This matters more than it sounds: unfenced retrieval on an older claim buries the primary report under everything published since.
  3. Check each claim. zetesis_scope, then zetesis_evidence with short entity-anchored queries, then grade against what came back. Work through them in order and report as you go rather than saving everything for the end. If a claim concerns funding, competitive position, or a company's standing, append the entity-only terms to queries and set include_capital: true; it is off by default and without it no grant or filing id can appear.
  4. Report per claim, briefly:
    • the claim, quoted or tightly paraphrased
    • whether the retrieved evidence supports it, partly supports it, or does not reach it
    • the hard ids that bear on it
    • if it fails, what the evidence actually says instead

Rules

Cite only ids that came back in a bundle. If nothing relevant was retrieved for a claim, say exactly that rather than filling the gap from memory; "no retrieved evidence bears on this" is a legitimate and useful result.

Do not rewrite the user's document unless they ask. The output is an assessment.

Close with the claims that need the author's attention first, ordered by how much the argument depends on them.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free