Files
gadfly/cmd/gadfly/specialists_test.go
T
Steve Dudenhoeffer 7809d1b93d
Build & push image / build-and-push (push) Successful in 8s
feat: specialist suite — configurable + custom review lenses (one consolidated comment)
Replace the single generic review with a suite of focused specialists, each its
own review+recheck pass, merged into ONE comment (a collapsible section per lens,
led by the worst verdict; the optional `improvements` lens never escalates it).

- cmd/gadfly/specialists.go: built-in lenses + default suite (security, correctness,
  maintainability, performance, error-handling) + opt-in (tests, docs, conventions,
  improvements). Selection via GADFLY_SPECIALISTS (csv/"all"); custom defs via
  GADFLY_SPECIALIST_<NAME> env and a repo .gadfly.yml (specialists + define).
  Precedence: built-ins < file < env. Unknown names error but don't sink the run.
- cmd/gadfly/consolidate.go: verdict parse + one-comment render.
- main.go: loop specialists; per-lens failure is an inline notice, never fatal.
  Default timeout bumped to 600s (suite runs sequentially).
- base system prompt trimmed to persona+tools+discipline+output; lens-specific
  focus is appended per specialist (semantic re-derivation discipline kept in base).
- entrypoint default models -> single model (suite already gives breadth; cost ~=
  specialists × models × 2). Adds gopkg.in/yaml.v3.
- docs/examples: README "Specialists" section, examples/.gadfly.yml, stub var,
  CLAUDE.md architecture/config. Dynamic `auto` selection is the planned next step.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-25 19:23:05 -04:00

119 lines
3.3 KiB
Go

package main
import (
"os"
"path/filepath"
"testing"
)
func names(specs []Specialist) []string {
out := make([]string, len(specs))
for i, s := range specs {
out[i] = s.Name
}
return out
}
func eq(a, b []string) bool {
if len(a) != len(b) {
return false
}
for i := range a {
if a[i] != b[i] {
return false
}
}
return true
}
func TestResolveSpecialists_DefaultSuite(t *testing.T) {
t.Setenv("GADFLY_SPECIALISTS", "")
specs, errs := resolveSpecialists(t.TempDir())
if len(errs) != 0 {
t.Fatalf("unexpected errors: %v", errs)
}
if !eq(names(specs), defaultSuite) {
t.Errorf("default = %v, want %v", names(specs), defaultSuite)
}
}
func TestResolveSpecialists_EnvSelection(t *testing.T) {
t.Setenv("GADFLY_SPECIALISTS", "security, tests")
specs, errs := resolveSpecialists(t.TempDir())
if len(errs) != 0 {
t.Fatalf("unexpected errors: %v", errs)
}
if !eq(names(specs), []string{"security", "tests"}) {
t.Errorf("got %v", names(specs))
}
}
func TestResolveSpecialists_UnknownNameErrors(t *testing.T) {
t.Setenv("GADFLY_SPECIALISTS", "security,bogus")
specs, errs := resolveSpecialists(t.TempDir())
if len(errs) == 0 {
t.Fatal("expected an error for unknown specialist")
}
if !eq(names(specs), []string{"security"}) {
t.Errorf("valid ones should still resolve, got %v", names(specs))
}
}
func TestResolveSpecialists_EnvCustomDefinition(t *testing.T) {
t.Setenv("GADFLY_SPECIALIST_MIGRATIONS", "Review DB migrations for destructive ops.")
t.Setenv("GADFLY_SPECIALISTS", "migrations")
specs, errs := resolveSpecialists(t.TempDir())
if len(errs) != 0 {
t.Fatalf("unexpected errors: %v", errs)
}
if len(specs) != 1 || specs[0].Name != "migrations" || specs[0].Focus == "" {
t.Fatalf("custom specialist not registered: %+v", specs)
}
}
func TestResolveSpecialists_FileConfig(t *testing.T) {
dir := t.TempDir()
cfg := `specialists: [security, migrations]
define:
- name: migrations
title: "DB migrations"
focus: "Review schema migrations for destructive or unindexed changes."
`
if err := os.WriteFile(filepath.Join(dir, ".gadfly.yml"), []byte(cfg), 0o644); err != nil {
t.Fatal(err)
}
t.Setenv("GADFLY_SPECIALISTS", "") // let the file drive selection
specs, errs := resolveSpecialists(dir)
if len(errs) != 0 {
t.Fatalf("unexpected errors: %v", errs)
}
if !eq(names(specs), []string{"security", "migrations"}) {
t.Errorf("got %v", names(specs))
}
}
func TestParseVerdictAndWorst(t *testing.T) {
cases := map[string]verdict{
"VERDICT: No material issues found.": verdictClean,
"Minor issues\n- nit": verdictMinor,
"**Blocking issues found**": verdictBlocking,
"something unparseable": verdictUnknown,
}
for in, want := range cases {
if got := parseVerdict(in); got != want {
t.Errorf("parseVerdict(%q) = %v, want %v", in, got, want)
}
}
results := []specialistResult{
{spec: Specialist{Name: "security"}, verdict: verdictMinor},
{spec: Specialist{Name: "correctness"}, verdict: verdictBlocking},
{spec: Specialist{Name: "improvements"}, verdict: verdictBlocking}, // must not count
}
if w := worstVerdict(results[:2]); w != verdictBlocking {
t.Errorf("worst = %v, want blocking", w)
}
if w := worstVerdict([]specialistResult{results[0], results[2]}); w != verdictMinor {
t.Errorf("improvements should not escalate; worst = %v, want minor", w)
}
}