346 lines
10 KiB
TypeScript
346 lines
10 KiB
TypeScript
import fs from "node:fs";
|
|
import path from "node:path";
|
|
import { fileURLToPath } from "node:url";
|
|
import { describe, expect, it } from "vitest";
|
|
import {
|
|
analyzeFrontmatterBlock,
|
|
detectFrontmatterRoundTripIssues,
|
|
getSkillFrontmatterUnknownKeys,
|
|
joinFrontmatterBlock,
|
|
parseFrontmatterFields,
|
|
parseFrontmatterMarkdown,
|
|
skillFrontmatterSchema,
|
|
splitFrontmatterBlock,
|
|
stringifyFrontmatter,
|
|
} from "./frontmatter.js";
|
|
|
|
const repoRoot = path.resolve(path.dirname(fileURLToPath(import.meta.url)), "../../..");
|
|
const skillMarkdownSearchRoots = [
|
|
"packages/skills-catalog/catalog",
|
|
"packages/teams-catalog/catalog",
|
|
"packages/adapters/hermes/skills",
|
|
"packages/plugins/plugin-llm-wiki/skills",
|
|
"skills",
|
|
];
|
|
|
|
describe("parseFrontmatterMarkdown", () => {
|
|
it("parses folded and literal YAML block scalars", () => {
|
|
const folded = parseFrontmatterMarkdown([
|
|
"---",
|
|
"name: Folded",
|
|
"description: >",
|
|
" First line",
|
|
" second line",
|
|
"",
|
|
" Third paragraph",
|
|
"---",
|
|
"",
|
|
"Body",
|
|
].join("\n"));
|
|
|
|
expect(folded.frontmatter.description).toBe("First line second line\n\nThird paragraph\n");
|
|
|
|
const literal = parseFrontmatterMarkdown([
|
|
"---",
|
|
"name: Literal",
|
|
"description: |",
|
|
" First line",
|
|
" second line",
|
|
"---",
|
|
"",
|
|
"Body",
|
|
].join("\n"));
|
|
|
|
expect(literal.frontmatter.description).toBe("First line\nsecond line\n");
|
|
});
|
|
|
|
it("respects block-scalar chomping indicators", () => {
|
|
const foldedStrip = parseFrontmatterMarkdown([
|
|
"---",
|
|
"description: >-",
|
|
" First line",
|
|
" second line",
|
|
"",
|
|
" Third paragraph",
|
|
"---",
|
|
"",
|
|
"Body",
|
|
].join("\n"));
|
|
|
|
expect(foldedStrip.frontmatter.description).toBe("First line second line\n\nThird paragraph");
|
|
|
|
const literalKeep = parseFrontmatterMarkdown([
|
|
"---",
|
|
"description: |+",
|
|
" First line",
|
|
" second line",
|
|
"",
|
|
"",
|
|
"---",
|
|
"",
|
|
"Body",
|
|
].join("\n"));
|
|
|
|
expect(literalKeep.frontmatter.description).toBe("First line\nsecond line\n\n");
|
|
});
|
|
|
|
it("parses inline object array items nested under frontmatter keys", () => {
|
|
const parsed = parseFrontmatterMarkdown([
|
|
"---",
|
|
"metadata:",
|
|
" sources:",
|
|
" - kind: github-dir",
|
|
" repo: paperclipai/paperclip",
|
|
" path: skills/paperclip",
|
|
"---",
|
|
"",
|
|
"Body",
|
|
].join("\n"));
|
|
|
|
expect(parsed.frontmatter).toMatchObject({
|
|
metadata: {
|
|
sources: [
|
|
{
|
|
kind: "github-dir",
|
|
repo: "paperclipai/paperclip",
|
|
path: "skills/paperclip",
|
|
},
|
|
],
|
|
},
|
|
});
|
|
});
|
|
|
|
it("does not treat trailing-dot decimals as numbers", () => {
|
|
const parsed = parseFrontmatterMarkdown([
|
|
"---",
|
|
"version: 1.",
|
|
"---",
|
|
"",
|
|
].join("\n"));
|
|
|
|
expect(parsed.frontmatter.version).toBe("1.");
|
|
});
|
|
});
|
|
|
|
describe("splitFrontmatterBlock", () => {
|
|
it("splits every bundled skill markdown file without losing bytes", () => {
|
|
const skillMarkdownFiles = collectSkillMarkdownFiles();
|
|
|
|
expect(skillMarkdownFiles.length).toBeGreaterThan(0);
|
|
for (const filePath of skillMarkdownFiles) {
|
|
const raw = fs.readFileSync(filePath, "utf8");
|
|
const split = splitFrontmatterBlock(raw);
|
|
const joined = split.hasFrontmatter
|
|
? `---\n${split.frontmatterText}\n---\n${split.body}`
|
|
: split.body;
|
|
|
|
expect(joined, path.relative(repoRoot, filePath)).toBe(raw);
|
|
}
|
|
});
|
|
|
|
it("leaves files without frontmatter untouched", () => {
|
|
const raw = "Body starts immediately.\n\n---\nThis is not frontmatter.\n";
|
|
const split = splitFrontmatterBlock(raw);
|
|
|
|
expect(split).toEqual({
|
|
frontmatterText: "",
|
|
body: raw,
|
|
hasFrontmatter: false,
|
|
});
|
|
});
|
|
|
|
it("treats an empty opening block as frontmatter", () => {
|
|
const raw = "---\n---\nBody\n";
|
|
|
|
expect(splitFrontmatterBlock(raw)).toEqual({
|
|
frontmatterText: "",
|
|
body: "Body\n",
|
|
hasFrontmatter: true,
|
|
});
|
|
});
|
|
});
|
|
|
|
describe("stringifyFrontmatter", () => {
|
|
it.each([
|
|
{
|
|
label: "nested metadata",
|
|
value: {
|
|
name: "demo-skill",
|
|
description: "Demo skill",
|
|
metadata: {
|
|
source: {
|
|
kind: "github-dir",
|
|
repo: "paperclipai/paperclip",
|
|
path: "skills/paperclip",
|
|
},
|
|
},
|
|
},
|
|
},
|
|
{
|
|
label: "arrays",
|
|
value: {
|
|
name: "tool-skill",
|
|
description: "Tool skill",
|
|
"allowed-tools": ["Read", "Write", "Bash"],
|
|
tags: ["skills", "frontmatter"],
|
|
},
|
|
},
|
|
{
|
|
label: "block scalars",
|
|
value: {
|
|
name: "block-skill",
|
|
description: "First line\nsecond line\n\nThird paragraph\n",
|
|
metadata: {
|
|
notes: "Keep\nall\nline breaks",
|
|
},
|
|
},
|
|
},
|
|
])("serializes parser-compatible YAML for $label", ({ value }) => {
|
|
const first = parseFrontmatterMarkdown(`---\n${stringifyFrontmatter(value)}\n---\n`).frontmatter;
|
|
const second = parseFrontmatterMarkdown(`---\n${stringifyFrontmatter(first)}\n---\n`).frontmatter;
|
|
|
|
expect(second).toEqual(first);
|
|
});
|
|
});
|
|
|
|
describe("skillFrontmatterSchema", () => {
|
|
it("validates core skill frontmatter fields while allowing unknown keys", () => {
|
|
const parsed = skillFrontmatterSchema.parse({
|
|
name: "demo-skill",
|
|
description: "A demo skill.",
|
|
"allowed-tools": ["Read", "Write"],
|
|
metadata: { nested: { enabled: true } },
|
|
tags: ["demo"],
|
|
});
|
|
|
|
expect(parsed.tags).toEqual(["demo"]);
|
|
expect(getSkillFrontmatterUnknownKeys(parsed)).toEqual(["tags"]);
|
|
});
|
|
|
|
it("rejects non-slug skill names", () => {
|
|
expect(() => skillFrontmatterSchema.parse({
|
|
name: "Demo Skill",
|
|
description: "A demo skill.",
|
|
})).toThrow();
|
|
});
|
|
});
|
|
|
|
describe("detectFrontmatterRoundTripIssues", () => {
|
|
it("reports YAML constructs that fields mode cannot preserve", () => {
|
|
const issues = detectFrontmatterRoundTripIssues([
|
|
"# leading comment",
|
|
"\"quoted-key\": value",
|
|
"base: &base",
|
|
"copy: *base",
|
|
].join("\n"));
|
|
|
|
expect(issues.map((issue) => issue.kind)).toEqual([
|
|
"comment",
|
|
"quoted_key",
|
|
"anchor",
|
|
"alias",
|
|
]);
|
|
});
|
|
});
|
|
|
|
describe("joinFrontmatterBlock", () => {
|
|
it("is the exact inverse of splitFrontmatterBlock (byte-identity round-trip)", () => {
|
|
const samples = [
|
|
"---\nname: reflection-coach\ndescription: A coach\n---\n# Body\n\nHello\n",
|
|
"---\nname: x\n---\nno trailing newline",
|
|
"---\nname: x\n---\n", // empty body
|
|
"---\ndescription: >\n folded\n text\n---\nBody with comment: value\n",
|
|
"# just markdown, no frontmatter\n",
|
|
"---\nunterminated frontmatter\nstill body",
|
|
"---\nmetadata:\n author: Paperclip\n # comment stays\n---\nbody\n",
|
|
];
|
|
for (const raw of samples) {
|
|
expect(joinFrontmatterBlock(splitFrontmatterBlock(raw))).toBe(raw);
|
|
}
|
|
});
|
|
|
|
it("returns the body untouched when there is no frontmatter", () => {
|
|
expect(
|
|
joinFrontmatterBlock({ frontmatterText: "", body: "just body", hasFrontmatter: false }),
|
|
).toBe("just body");
|
|
});
|
|
});
|
|
|
|
describe("parseFrontmatterFields", () => {
|
|
it("parses the raw block text into an object and is lenient on garbage", () => {
|
|
expect(parseFrontmatterFields("name: foo\ndescription: bar")).toEqual({
|
|
name: "foo",
|
|
description: "bar",
|
|
});
|
|
expect(parseFrontmatterFields("")).toEqual({});
|
|
expect(parseFrontmatterFields("# only a comment")).toEqual({});
|
|
});
|
|
});
|
|
|
|
describe("analyzeFrontmatterBlock", () => {
|
|
it("marks a simple inline block as round-trippable", () => {
|
|
const result = analyzeFrontmatterBlock("name: reflection-coach\ndescription: A coach");
|
|
expect(result.canRoundTrip).toBe(true);
|
|
expect(result.issues).toEqual([]);
|
|
expect(result.parsed).toEqual({ name: "reflection-coach", description: "A coach" });
|
|
});
|
|
|
|
it("marks a block with allowed-tools and metadata as round-trippable", () => {
|
|
const raw = [
|
|
"name: coach",
|
|
"description: A coach",
|
|
"allowed-tools:",
|
|
" - Read",
|
|
" - Grep",
|
|
"metadata:",
|
|
" author: Paperclip",
|
|
" version: 2",
|
|
].join("\n");
|
|
const result = analyzeFrontmatterBlock(raw);
|
|
expect(result.canRoundTrip).toBe(true);
|
|
expect(result.parsed["allowed-tools"]).toEqual(["Read", "Grep"]);
|
|
});
|
|
|
|
it("refuses fields mode when comments are present (would be dropped)", () => {
|
|
const result = analyzeFrontmatterBlock("name: coach # inline note\ndescription: x");
|
|
expect(result.canRoundTrip).toBe(false);
|
|
expect(result.issues.some((issue) => issue.kind === "comment")).toBe(true);
|
|
});
|
|
|
|
it("refuses fields mode for folded scalars the serializer cannot reproduce", () => {
|
|
const raw = ["description: >", " first line", " second line"].join("\n");
|
|
const result = analyzeFrontmatterBlock(raw);
|
|
// No detector "issue", but re-serialization is not byte-identical, so it is
|
|
// still not round-trippable — the strict serialize-back gate catches it.
|
|
expect(result.canRoundTrip).toBe(false);
|
|
});
|
|
|
|
it("treats an empty block as round-trippable", () => {
|
|
const result = analyzeFrontmatterBlock("");
|
|
expect(result.canRoundTrip).toBe(true);
|
|
expect(result.parsed).toEqual({});
|
|
});
|
|
});
|
|
|
|
function collectSkillMarkdownFiles() {
|
|
return skillMarkdownSearchRoots.flatMap((relativeRoot) => {
|
|
const absoluteRoot = path.join(repoRoot, relativeRoot);
|
|
return fs.existsSync(absoluteRoot) ? collectSkillMarkdownFilesUnder(absoluteRoot) : [];
|
|
}).sort();
|
|
}
|
|
|
|
function collectSkillMarkdownFilesUnder(root: string): string[] {
|
|
const files: string[] = [];
|
|
for (const entry of fs.readdirSync(root, { withFileTypes: true })) {
|
|
const absolutePath = path.join(root, entry.name);
|
|
if (entry.isDirectory()) {
|
|
files.push(...collectSkillMarkdownFilesUnder(absolutePath));
|
|
continue;
|
|
}
|
|
if (entry.isFile() && entry.name === "SKILL.md") {
|
|
files.push(absolutePath);
|
|
}
|
|
}
|
|
return files;
|
|
}
|