agentleFS
vcxcviivcxcvii/rainmaker/skills/can-i-actually-rank/SKILL.md

can-i-actually-rank

Skill0 starsChanged 4 months agoMITView on GitHub

Check whether a candidate keyword is actually winnable by reading the live SERP, then apply five commercial filters. Ends every candidate in QUALIFY, CONDITIONAL or KILL, kills freely, and never treats vendor authority scores as evidence.

  • Reads credentials

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 4 are distinctive, and 6 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 6

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Run rainmaker context --check first.

    line 24
  2. Very specific:

    Only the data/ files listed in this skill's Consumes table.

    line 32
  3. Very specific:

    If strategy.json.context_hash does not match the current hash of context/business.md, say exactly:

    line 34
  4. Very specific:

    If context/business.md carries confidence: stub, continue, and stamp every output with: "Built on a stub context.

    line 40
  5. Very specific:

    After approval, use rainmaker serp --allow-paid "<query>".

    line 60
  6. Very specific:

    Read the People Also Ask and related searches from the capture when present.

    line 66
  7. Distinctive:

    Pipeline, not traffic: If this ranks and drives traffic, will any of those readers plausibly buy?

    line 68
  8. Distinctive:

    Never verdict from volume and difficulty data alone.

    line 78
  9. Distinctive:

    Never cite a third-party domain authority score anywhere in this skill's reasoning or output.

    line 82
  10. Distinctive:

    Every candidate has a verdict traceable to a real SERP capture, every CONDITIONAL carries both a condition and a resolution path, every …

    line 107

≈ “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.

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

can-i-actually-rank

Context load

Run rainmaker context --check first. It prints what exists, what is stale, and exits 1 if anything this skill requires is missing.

Then read, in this order:

  1. context/business.md in full. If absent, stop: "No business context. Run rainmaker audit, then the know-my-buyer skill."
  2. context/voice.md if this skill writes prose. If absent, stop and say so.
  3. context/glossary.md if this skill names products, features or competitors.
  4. data/strategy.json if this skill reads or writes strategy.
  5. Only the data/ files listed in this skill's Consumes table. Never crawl or call an API the core already covers.

If strategy.json.context_hash does not match the current hash of context/business.md, say exactly:

"Business context was edited after the strategy was written. Re-run know-my-buyer, or run rainmaker context --sync to accept the prose as authoritative."

Then stop.

If context/business.md carries confidence: stub, continue, and stamp every output with: "Built on a stub context. Nothing in it came from a buyer. Run know-my-buyer to replace it."

Consumes

FileWhyIf missing
candidate queries from pick-my-battles or what-to-target-nextwhat to checkask the user for a query list
data/snapshots/<latest>/serp.jsonthe live SERP capturerun rainmaker serp <query> first for each candidate
data/snapshots/<latest>/gsc.jsonour own demonstrated ceiling at similar volumecontinue without the ceiling signal
context/glossary.mdcategory terms, for the category-match checkcontinue, category check assumes present

Produces

keyword_plan[].verdict (this skill), and blueprint.nodes[].serp_verdict once a blueprint exists.

Refuses when

No SERP capture exists for a candidate. Say: "No SERP capture for '<query>'. Ask me to approve Firecrawl credits, then run rainmaker serp --allow-paid \"<query>\" and re-run this skill." Do not guess a verdict from Ahrefs-style volume data alone; the whole point of this skill is reading the actual SERP.

rainmaker serp spends Firecrawl credits. Ask for approval before running it. After approval, use rainmaker serp --allow-paid "<query>". Never infer approval from an ambient FIRECRAWL_API_KEY.

Procedure

  1. For each candidate, load its serp.json capture and call the verdict engine (src/serp/verdict.ts computeVerdict), which is the tested, deterministic half of this skill. It returns intent consistency, category presence, the rewarded format, beatability evidence, and a verdict of QUALIFY, CONDITIONAL or KILL.
  2. Read the People Also Ask and related searches from the capture when present; feed them into cluster completeness in map-my-site.
  3. For every survivor (QUALIFY or CONDITIONAL), apply the five commercial filters, in order, against context/business.md:
    • Pipeline, not traffic. If this ranks and drives traffic, will any of those readers plausibly buy? Fail if the keyword attracts an audience the product does not serve, or the traffic signal comes from geographies the site does not sell into.
    • ICP match. Does the implied searcher match the ICP on two or more of role, industry, size, problem? Reference strategy.json.icp and personas.
    • Honest product fit. Can the product genuinely solve the implied problem without a claim it cannot honestly support? Check against "What we will not say" in business.md.
    • Alignment. Does the recommendation contradict anything the client has explicitly ruled out?
    • Net new or overlap. Does an existing URL already target this intent? If so, this is a refresh or consolidation, never a new page; check blueprint.json and crawl.json before proposing anything new.
  4. A candidate failing any filter moves from QUALIFY or CONDITIONAL to KILL, with the failing filter named as the reason.
  5. Summarise the two or three kill patterns that eliminated the most candidates. If half the pool died on "wrong category dominates the SERP", that is a positioning problem upstream of content, and it belongs in the summary, not buried in a per-row note.

Decision rules

  • Never verdict from volume and difficulty data alone. Read the actual SERP capture every time.
  • Kill freely. This skill exists to reduce the candidate pool to only winnable, correctly targeted keywords; keeping marginal ones wastes briefing, writing and publishing effort on targets that will not rank or will not convert.
  • "Beatable" requires the named evidence the verdict engine already computed: a lower-page-count competitor, an intent-mismatched result, this site's own demonstrated ceiling, a format gap, or stale top results. Optimism is not evidence and is never added as a sixth reason.
  • A CONDITIONAL verdict always carries condition and condition_resolved_by. A condition that cannot be resolved is reclassified as KILL, not left as a permanent soft pass.
  • Never cite a third-party domain authority score anywhere in this skill's reasoning or output.

Output

## SERP qualification

Checked: <n>   QUALIFY: <n>   CONDITIONAL: <n>   KILL: <n>

<query>: QUALIFY
  rewarded format: <format>   evidence: <reason>, <reason>

<query>: CONDITIONAL
  condition: <text>
  resolved by: <text>

<query>: KILL
  reason: <text>

Most common kill reasons:
  <reason>: <n> candidates

Done when

Every candidate has a verdict traceable to a real SERP capture, every CONDITIONAL carries both a condition and a resolution path, every KILL states which filter or which missing evidence eliminated it, and the summary names the dominant kill pattern.

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.