M2: skill index — SKILL.md parsing, pgvector index, query by role

Skills module (references SharedKernel only):
- Skill entity + SkillsDbContext (schema "skills") + InitialSkills migration: roles/tools/
  context as text[], risk-tagged actions and golden tests as jsonb, a nullable vector(384)
  embedding, unique (SkillKey, Version).
- SkillMarkdownParser: YAML frontmatter (YamlDotNet) + markdown body → SkillManifest.
- HashingSkillEmbedder: placeholder deterministic embedder so the pgvector path is real now;
  swapped for ONNX/BYOK embeddings at M3-M4 (384-dim to match MiniLM/bge).
- SkillIndexer: parse → hash → embed → upsert; structural publish gate (roles + >=1 golden
  test). Executing golden tests against a model + gating on edit distance lands at M4.
- Endpoints: GET /api/skills (filter by role/visibility), GET /api/skills/{key},
  POST /api/skills/index (manual/admin) — all authenticated.

Verified: build green; ArchitectureTests 8/8 (Skills references only SharedKernel);
IntegrationTests 21/21 incl. a new skill-registry flow — index a SKILL.md, it publishes,
is queryable by role (and not under others), re-index dedups, malformed is 400, catalogue
needs auth.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
soroush.asadi
2026-06-09 18:01:37 +03:30
parent ce5c644c7b
commit 401e3e69af
17 changed files with 1103 additions and 14 deletions
@@ -0,0 +1,109 @@
using System.Net;
using System.Net.Http.Headers;
using System.Net.Http.Json;
using Xunit;
namespace TeamUp.IntegrationTests;
/// <summary>
/// M2 skill-registry acceptance: index a SKILL.md, then find it queryable by role; a skill with
/// roles + golden tests publishes, a malformed one is rejected, and the catalogue needs auth.
/// </summary>
public sealed class SkillRegistryTests(PostgresFixture postgres) : IClassFixture<PostgresFixture>
{
private const string SpecWritingSkill =
"""
---
id: spec-writing
name: Spec Writing
version: 1.0.0
summary: Turn a feature request into a structured spec and child stories.
roles: [product-owner]
inputs: A feature request or task.
outputs: A spec plus proposed child stories.
actions:
- name: write-spec
risk: draft
- name: create-child-stories
risk: draft
tools: []
context: [house-style]
visibility: public
min_tier: free
golden_tests:
- input: "Add a logout button"
expected: "Spec: a logout button in the header that ends the session."
---
# Spec Writing
Write a clear, testable spec, then propose child stories.
""";
private sealed record BootstrapResponse(string Token, Guid MemberId, Guid OrganizationId);
private sealed record ActionDto(string Name, string Risk);
private sealed record SkillSummary(
string SkillKey, string Name, string Version, string? Summary, List<string> Roles,
string Visibility, string MinTier, string Status, List<ActionDto> Actions);
private sealed record SkillDetail(
SkillSummary Skill, string? Inputs, string? Outputs, List<string> Tools,
List<string> Context, int GoldenTestCount, string Body);
[Fact]
public async Task Index_publishes_and_makes_skill_queryable_by_role()
{
await using var factory = new TeamUpWebFactory(postgres.ConnectionString);
using var anon = factory.CreateClient();
// The catalogue requires auth.
Assert.Equal(HttpStatusCode.Unauthorized, (await anon.GetAsync("/api/skills/")).StatusCode);
var bootstrap = await anon.PostAsJsonAsync("/api/identity/bootstrap", new
{
organizationName = "AliaSaaS",
ownerEmail = "owner@alia.test",
ownerDisplayName = "Owner",
ownerPassword = "Passw0rd!",
});
var owner = await bootstrap.Content.ReadFromJsonAsync<BootstrapResponse>();
Assert.NotNull(owner);
using var client = factory.CreateClient();
client.DefaultRequestHeaders.Authorization = new AuthenticationHeaderValue("Bearer", owner!.Token);
// Index the SKILL.md.
var indexResponse = await client.PostAsJsonAsync("/api/skills/index", new { content = SpecWritingSkill });
Assert.Equal(HttpStatusCode.OK, indexResponse.StatusCode);
var indexed = await indexResponse.Content.ReadFromJsonAsync<SkillDetail>();
Assert.NotNull(indexed);
Assert.Equal("spec-writing", indexed!.Skill.SkillKey);
Assert.Equal("Published", indexed.Skill.Status); // has roles + a golden test
Assert.Equal(1, indexed.GoldenTestCount);
Assert.Contains(indexed.Skill.Actions, a => a.Name == "write-spec" && a.Risk == "Draft");
// Queryable by its role…
var forPo = await client.GetFromJsonAsync<List<SkillSummary>>("/api/skills/?role=product-owner");
Assert.Contains(forPo!, s => s.SkillKey == "spec-writing");
// …but not under an unrelated role.
var forQa = await client.GetFromJsonAsync<List<SkillSummary>>("/api/skills/?role=qa");
Assert.DoesNotContain(forQa!, s => s.SkillKey == "spec-writing");
// Detail by key.
var detail = await client.GetFromJsonAsync<List<SkillDetail>>("/api/skills/spec-writing");
Assert.Single(detail!);
Assert.Equal("public", detail![0].Skill.Visibility.ToLowerInvariant());
// Re-indexing the same key+version updates in place (no duplicate).
await client.PostAsJsonAsync("/api/skills/index", new { content = SpecWritingSkill });
var afterReindex = await client.GetFromJsonAsync<List<SkillDetail>>("/api/skills/spec-writing");
Assert.Single(afterReindex!);
// A malformed SKILL.md (no frontmatter) is rejected.
var bad = await client.PostAsJsonAsync("/api/skills/index", new { content = "# no frontmatter here" });
Assert.Equal(HttpStatusCode.BadRequest, bad.StatusCode);
}
}