test: migrate all Go tests to testify

[wesm]

Summary

Standardizes Go tests on github.com/stretchr/testify. All 172 test files now use assert.X / require.X instead of t.Errorf / t.Fatalf.

What changed

• t.Fatal* mapped to require.X (halts on failure).
• t.Error* mapped to assert.X (continues on failure).
• Equality argument order is (want, got) per testify convention.
• Custom helpers in internal/testutil/assert.go (AssertEqual, MustNoErr, etc.) removed; MakeSet kept.
• internal/testutil/email/email.go::AssertStringSliceEqual removed (unused after migration).
• testify promoted to a direct dependency in go.mod.
• CLAUDE.md gains a ## Testing section documenting the convention.
• New AGENTS.md at the repo root reiterates the same rule for any AI coding agent.

Coverage

Production-code coverage is unchanged. Total moves 56.8% → 57.0% (improvement comes from testutil helpers whose error branches are no longer uncovered). 46/46 packages pass; same set as before.

How to use

New tests:

import (
    "github.com/stretchr/testify/assert"
    "github.com/stretchr/testify/require"
)

require.NoError(t, err, "context")
assert.Equal(t, want, got)

Build tags fts5 sqlite_vec are still required for go test; make test sets them.

[roborev-ci]

roborev: Combined Review (ccec23e)

No Medium, High, or Critical issues found across the reviews.

---

Synthesized from 3 reviews (agents: codex, gemini | types: default, security)

[roborev-ci]

roborev: Combined Review (9ab93d7)

No Medium, High, or Critical findings were reported.

Both substantive review outputs indicate no issues found; the remaining review outputs are empty.

---

Synthesized from 3 reviews (agents: codex, gemini | types: default, security)

[mariusvniekerk]

@wesm want too copy over the structure enforcement tool from middleman ?

[wesm]

yep, clanking

[wesm]

rebasing

[roborev-ci]

roborev: Combined Review (8df410f)

Verdict: No medium-or-higher issues were found across the reviews.

All agents reported the code is clean or provided no findings.

---

Synthesized from 3 reviews (agents: codex, gemini | types: default, security)

[roborev-ci]

roborev: Combined Review (4f35c37)

Summary verdict: No Medium, High, or Critical issues were reported.

All completed review agents found the code clean.

---

Synthesized from 3 reviews (agents: codex, gemini | types: default, security)

Note: gemini review skipped (agent quota exhausted)

[wesm]

Merging. The CI failure is unrelated