agentleFS
ansvisoransvisor/ansvisor/skills/facts/SKILL.md

facts

Skill233 starsChanged 5 months agoMITView on GitHub

Checks every draft against your approved brand facts.

5 other projects use a version of this, 1 of them edited.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 5 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 5

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Check the brand's drafts against its approved brand facts (its plans, trial, founding year and the claims it makes), and list every place a …

    line 12
  2. Very specific:

    Read the approved facts with get_brand_facts.

    line 32
  3. Distinctive:

    Read each draft and find every passage that states one of those facts: a plan, a trial, a founding year, a claim about the brand.

    line 33
  4. Distinctive:

    Don't decide a fact is true from your own knowledge or from a page you read: only an approved fact, or one the person confirms, counts.

    line 47
  5. Distinctive:

    Don't change anything in a passage but the fact.

    line 48
  6. Distinctive:

    Keep the passages AI quotes word for word unless the fact in them is wrong, and then say they are quoted.

    line 49

“Common” is measured: kept by at least 6 in 10 of the 4 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

Brand facts check

Check the brand's drafts against its approved brand facts (its plans, trial, founding year and the claims it makes), and list every place a draft says something else.

When to use

  • "Check our drafts against our brand facts", "Does this page still say the old trial length?", "Are our facts consistent?"

Where the facts come from

Ansvisor reads the brand facts from the brand's own pricing, About and product pages, and a person approves each one in the brand's settings, on the Brand facts tab. Only approved facts count. Read them with get_brand_facts: the approved facts still on the brand's site, each in its page's own words, with when it was approved and the page it came from. Facts waiting for approval, rejected ones and ones gone from the site aren't in it.

When it returns none, say so, and ask the person to open Brand facts and approve the ones waiting, or to paste the facts they confirm. Never approve a fact yourself, and never use a fact that isn't approved or that the person didn't confirm.

What to read

  1. get_brand_facts: the brand's approved facts.
  2. list_drafts: the brand's drafts, latest changed first.
  3. get_draft for each draft you check, at most five in one run: its passages, and where it is live.

Steps

  1. Read the approved facts with get_brand_facts. With none, stop and ask for them, as above. Facts the person gives in the chat count too.
  2. Read each draft and find every passage that states one of those facts: a plan, a trial, a founding year, a claim about the brand.
  3. Compare each with the approved fact: same, different, or no longer true.
  4. Draft the fix for each mismatch: the passage with the approved fact in it, nothing else changed.

Output

  1. A table: Draft, Passage, Says, Approved fact, Fix.
  2. The drafts with no mismatch, in one line.
  3. Facts a draft states that aren't in the approved list, for the person to check.

When the person wants the fixes kept, save the corrected page as a new draft with create_draft; the app asks them first when it should. In an agent's run nobody is there to ask: always save it with its tool, and the app holds it for approval where the agent's mode needs it.

Don't

  • Don't decide a fact is true from your own knowledge or from a page you read: only an approved fact, or one the person confirms, counts.
  • Don't change anything in a passage but the fact.
  • Keep the passages AI quotes word for word unless the fact in them is wrong, and then say they are quoted.

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.