agentleFS
bm629bm629/agent-skills/skills/reviewing-integrations-prior-art-survey/SKILL.md

reviewing-integrations-prior-art-survey

Skill2 starsChanged 3 months agoMITView on GitHub

Use when reviewing an integrations prior-art artifact — an integration vocabulary map, one angle's search output, one service's extract record, or the integration register and its report — before it is accepted. Judges what a deterministic gate cannot: whether a locator host really is the vendor's own, whether an evidence quote supports the claim drawn from it, whether an authority band is defensible for the page it points at, whether a record was deep-read first-party or restated from a catalog, whether each complexity component is the one the facts support, whether a lens denominator counts what was REACHED, and whether an admission or an absence was recorded truthfully. Emits exactly one verdict, approve or revise, with every finding tied to a numbered condition. Keywords: integrations prior art review, connector catalog review, descriptor evidence, vendor scope, coverage grid, integration register, prior-art reviewer.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    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:

    Do not re-check what the validator refuses: Every condition below names the rule that owns the other half where one exists, and a finding …

    line 29
  2. Very specific:

    Read references/conditions.md.

    line 62

≈ “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

Reviewing an integrations prior-art artifact

You are the second half of a two-part gate. The first half already ran.

Do not re-check what the validator refuses. Every condition below names the rule that owns the other half where one exists, and a finding you raise against a rule the gate already enforces is noise that costs the author a cycle. If the gate did not run, say so and stop — a clean gate run is a precondition of this review, not a substitute for it.

Your evidence

sourcewhat it settles
the artifact under revieweverything below
the vocabulary map it was produced againstwhether the owed grid and the angle verdicts are right
the SCOPE and CLASSIFICATION the producer was handedwhether meta.classification is a faithful transcription
integrations-prior-art-survey/schemas/*.jsonwhat each field is FOR, in its description
integrations-prior-art-survey/references/source-registry.yamlcaps, orderings, bands, complete_listing, the excluded block
integrations-prior-art-survey/references/angles/<angle>.mdthat angle's mechanism, sources and precondition
integrations-prior-art-survey/references/absent-input-policy.mdthe nine catalog auth modes and which of them map to null — C9
integrations-prior-art-survey/references/category-vocabulary.mdthe seeded category vocabulary — C13
integrations-prior-art-survey/references/integration-vocabulary-map-guide.mdthe probe's three checks — C20
the FROZEN extract-queue.yaml, and the extract records themselveswhat extraction was asked to produce, and what it wrote — C26, C28
integrations-prior-art-survey/references/extraction-template-guide.mdwhat each record field is FOR, field by field — C21-C25
integrations-prior-art-survey/references/synthesis-lenses.mdthe eight formulas and the denominator each divides by — C26-C29
integrations-prior-art-survey/references/synthesis-report-guide.mdthe report's eight fixed sections, and what does not go in it — C30

THIRTEEN sources, and NINE of them are PRODUCER-package paths you read rather than files this package ships -- the schemas, the source registry, the angle references, the absent-input policy, the category vocabulary, the vocabulary-map guide, the extraction template guide, the synthesis lenses and the report guide. All nine. A review that reads only the artifact and the map can settle the arithmetic and none of the vocabulary. The scope and classification are on the list because meta.classification is a TRANSCRIPTION, and judged only against itself a fabricated value reads exactly like a real one.

Conditions

Read references/conditions.md. Judge every condition that applies. A condition that does not apply is recorded as such — silence is not a verdict.

Your verdict

Emit exactly one line, last:

VERDICT: approve

or

VERDICT: revise

revise requires at least one finding naming its condition. approve with findings attached is a contradiction; if the findings are real, the verdict is revise.

Do not false-revise honest work. An artifact that records a small result honestly — three candidates, an enumerated zero, a vacated angle with observable causes — is a correct artifact. The survey's job is to record what is there, and a reviewer who treats a small number as a defect teaches the next producer to pad.

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.