tactic-check
Skill1 starsChanged 3 days agoMITView on GitHub
Check an AEO, GEO, or AI SEO tactic, an agency proposal, or a tool pitch against the platforms'' own rules and the evidence. Gives each tactic one result: against a rule, does not do what is claimed, unproven, or reasonable, with the official source. Use for "is this AEO tactic safe", "should I add llms.txt", "review this agency proposal", or "can I pay for reviews". Needs no Frontmost AI account.
In this repository18 other files it gives its agents.
- ai-readinessSkill
- ai-search-strategySkill
- answer-accuracySkill
- answer-pagesSkill
- community-repliesSkill
- customer-surveySkill
- directory-listingsSkill
- listicle-outreachSkill
- mention-checkSkill
- name-checkSkill
- page-draftingSkill
- placement-trackerSkill
- question-draftingSkill
- question-gapsSkill
- rival-gapsSkill
- scan-reviewSkill
- site-auditSkill
- source-mapSkill
The file
We split this skill into the rules it gives an agent
7 rules. About 0 are common, 7 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈7
Distinctive · estimate
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
- Distinctive: line 10
Use this guidance to check a tactic before the person pays for it or does it.
- Distinctive: line 14
Read the proposal, the pitch, or the plan.
- Distinctive: line 16
If a line is not clear enough to check, write the question that makes it clear.
- Distinctive: line 29
Start with tactics.md. It holds the results for the usual tactics, with the official sources.
- Distinctive: line 35
Use only these as a source for the result Against a rule: the platform's own policy page, or a regulator's page.
- Distinctive: line 41
A before-and-after result: Ask five things: which questions, how many answers for each question, which dates, which AI app, and what stayed …
- Distinctive: line 58
If the person already did a tactic that is against a rule, say how to undo it.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Tactic check
Many people sell tactics for AI search. Some tactics break a platform's rules, and the business carries the risk. Some have no evidence, and the business carries the cost.
Use this guidance to check a tactic before the person pays for it or does it.
1. List the tactics
Read the proposal, the pitch, or the plan. Write one line for each tactic: what is done, on which site, by whom, and at what cost.
If a line is not clear enough to check, write the question that makes it clear. Do not guess what the seller means.
2. Give each tactic one result
| Result | Rule |
| Against a rule | A platform's own published rule, or a law, forbids it. Quote the rule and link it |
| Does not do what is claimed | The platform's own documents say that the tactic has a different effect |
| Unproven | No rule forbids it. No platform document and no public test shows that it works |
| Reasonable | A platform's own guidance supports it, or it helps the people who read the page |
Some tactics depend on a condition. Give one result for each condition. A paid placement is reasonable when the page says that it is paid, and against a rule when the page hides it.
Start with tactics.md. It holds the results for the usual tactics, with the official sources.
3. Check the source today
Rules change. Before you quote a rule, open its official page. If you cannot open it, write "not checked today" next to the result.
Use only these as a source for the result Against a rule: the platform's own policy page, or a regulator's page. A blog post or a video is not a source for it.
4. Check the promised results
- A promise of mentions or of a place in ChatGPT. Nobody outside an AI company controls its answers. No method is proven to earn a mention. Give a promise the result Unproven.
- A number for how often people ask AI apps a question. The AI companies do not publish these numbers. Ask where the number comes from.
- A before-and-after result. Ask five things: which questions, how many answers for each question, which dates, which AI app, and what stayed unchanged for comparison. One answer before and one answer after show nothing.
5. Write the questions for the seller
For each tactic that is unproven or that does not do what is claimed, write the questions that fit, for the person to send:
- "Which platform rule or document supports this?"
- "How did you measure the result, and can I see the answers?"
- "Who owns the accounts, the pages, and the listings after the contract ends?"
- "Does a page that you pay for say that it is paid?"
Rules
- Give a result only with its reason. A result with no source is your opinion. Say so.
- This is not legal advice. For a question about the law, the person asks a lawyer.
- Laws differ by country. The reference file names United States rules. Say which country a rule covers.
- The decision is the person's. Report the risk and the cost. Do not decide for them.
- If the person already did a tactic that is against a rule, say how to undo it. Do not judge.
Output
A table with one row for each tactic: the tactic, the result, the reason, the source, and what to do in its place.
Then three lines:
- The tactics that are against a rule, with the risk of each one.
- The total cost of the tactics that are unproven or that do not do what is claimed.
- The questions for the seller.
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
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.

