Files
msd-core/tests/bug-2794-opencode-model-profile-overrides.test.cjs
Tom Boucher a28dcec981 chore(#597): replace count-based ratchet guards with AST lint + named-set allowlists (#603)
The windows-test-parity ratchet greps test source for fs.rmSync-without-
maxRetries (and six other Windows-portability anti-patterns), failing when an
integer offender COUNT exceeds a frozen baseline (rmSync: 95). A count ratchet
is a Goodhart metric: fixing one offender and adding another keeps the count
constant, so a new defect slips through green. Replace it — and every other
count ratchet in the repo — with a layered, masking-proof design.

Behavioral seam test
- tests/helpers-cleanup.test.cjs proves helpers.cleanup() carries the Windows
  EBUSY retry budget. cleanup() delegates retries to Node's fs.rmSync via
  maxRetries (it owns no loop), so the test asserts the option contract
  (recursive/force/maxRetries>0/retryDelay>0) + real-FS removal + the cwd-guard,
  rather than a loop that does not exist. The EBUSY risk is now tested ONCE at
  the helper, not approximated textually at every call site.

Write-time ESLint rule (AST-accurate, replaces the grep)
- eslint-rules/no-raw-rmsync-in-tests.cjs (error in tests/**/*.test.cjs) bans
  raw fs.rmSync, steering to cleanup(). Catches member, computed (fs['rmSync']),
  destructured and aliased forms; escape hatch is inline
  `// eslint-disable-next-line local/no-raw-rmsync-in-tests -- <reason>` only.
- Migrated 336 raw fs.rmSync teardown calls across ~116 test files to cleanup().
  ~18 genuinely load-bearing sites (mid-test SUT/fault-injection removals,
  error-swallowing or name-colliding local teardown helpers) keep the raw call
  with an inline eslint-disable + reason.

Shared anti-ratchet primitive
- scripts/lib/allowlist-ratchet.cjs:
  - assertWithinAllowlist: fails on NOVEL ids (new offender introduced) AND on
    STALE ids (a known offender was fixed but not pruned) — identity, not count,
    and a ratchet DOWN toward zero.
  - assertTightCeiling: a size/length budget whose ceiling must stay within a
    grace band of the high-water mark, so budgets may only tighten, never creep.

Ratchets converted onto the primitive
- windows-test-parity-guard.test.cjs: rmSync rule deleted (now ESLint-enforced);
  the remaining six patterns moved from integer baselines to named-set
  allowlists with ratchet-down.
- scripts/lint-test-file-count.{cjs,allowlist.json}: per-module integer counts →
  named filename sets (closes the swap-a-file-keep-the-count blind spot); a
  module dropping under cap now FAILS to force pruning its allowlist entry.
- enh-2790 skill-count `<= 63` → named skill allowlist (ratchets toward ~58).

Size budgets hardened (tighten-only)
- agent-size / workflow-size / feat-3039 help-tiered: ceilings lowered to the
  current high-water mark and an assertTightCeiling anti-creep check added per
  tier. Fixed external-contract limits (description ≤100 chars, agent ≤100 KB)
  are intentionally left as-is — they are not grandfathered creeping budgets.

No user-facing behavior change (tests + tooling only); no USER_FACING_PREFIXES
touched, so no changeset fragment is required.

Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-01 22:43:49 -04:00

240 lines
8.1 KiB
JavaScript

/**
* Regression test for bug #2794
*
* OpenCode generated agents ignored `model_profile_overrides.opencode.*`.
* The agent install path called `readGsdEffectiveModelOverrides` (explicit
* per-agent overrides) but never called `readGsdRuntimeProfileResolver`
* (tier-based profile overrides). When a user configured:
*
* { runtime: "opencode", model_profile_overrides: { opencode: { sonnet: "..." } } }
*
* generated `.opencode/agents/gsd-*.md` files contained no `model:` frontmatter.
*
* The fix adds a tier-resolver fallback in the OpenCode agent conversion block:
* explicit `model_overrides[agent]` > `model_profile_overrides.opencode.<tier>` > omit.
*
* This test exercises:
* 1. `readGsdRuntimeProfileResolver` correctly resolves OpenCode tier overrides.
* 2. The agent install code path embeds the resolved model into OpenCode frontmatter.
* 3. Explicit `model_overrides` still wins over tier-based resolution.
* 4. Missing overrides produce no `model:` field (no regression on omit behavior).
*/
process.env.GSD_TEST_MODE = '1';
const { describe, test, beforeEach, afterEach } = require('node:test');
const assert = require('node:assert/strict');
const fs = require('node:fs');
const path = require('node:path');
const os = require('node:os');
const {
readGsdRuntimeProfileResolver,
readGsdEffectiveModelOverrides,
convertClaudeToOpencodeFrontmatter,
install,
} = require('../bin/install.js');
const { createTempDir, cleanup } = require('./helpers.cjs');
const makeTmp = (prefix) => createTempDir(`gsd-2794-${prefix}-`);
function writeJson(p, obj) {
fs.mkdirSync(path.dirname(p), { recursive: true });
fs.writeFileSync(p, JSON.stringify(obj, null, 2), 'utf-8');
}
describe('bug-2794: readGsdRuntimeProfileResolver resolves opencode tier overrides', () => {
let projectDir;
let homeDir;
let origHome;
beforeEach(() => {
projectDir = makeTmp('proj');
homeDir = makeTmp('home');
origHome = process.env.HOME;
process.env.HOME = homeDir;
});
afterEach(() => {
if (origHome === undefined) delete process.env.HOME;
else process.env.HOME = origHome;
cleanup(projectDir);
cleanup(homeDir);
});
test('resolves opencode sonnet tier to user-supplied model ID', () => {
writeJson(path.join(projectDir, '.planning', 'config.json'), {
runtime: 'opencode',
model_profile: 'balanced',
model_profile_overrides: {
opencode: {
sonnet: 'anthropic/claude-sonnet-4-7',
},
},
});
const resolver = readGsdRuntimeProfileResolver(projectDir);
assert.ok(resolver !== null, 'expected a resolver for opencode runtime');
// gsd-roadmapper balanced tier = sonnet — should resolve to override
const entry = resolver.resolve('gsd-roadmapper');
assert.ok(entry !== null, 'expected entry for gsd-roadmapper');
assert.strictEqual(entry.model, 'anthropic/claude-sonnet-4-7', 'sonnet override applied');
});
test('returns null resolver when runtime is not set', () => {
writeJson(path.join(projectDir, '.planning', 'config.json'), {
model_profile: 'balanced',
model_profile_overrides: { opencode: { sonnet: 'x' } },
});
const resolver = readGsdRuntimeProfileResolver(projectDir);
assert.strictEqual(resolver, null, 'no resolver without runtime field');
});
test('resolver returns null for agent not in MODEL_PROFILES', () => {
writeJson(path.join(projectDir, '.planning', 'config.json'), {
runtime: 'opencode',
model_profile: 'balanced',
model_profile_overrides: { opencode: { sonnet: 'x' } },
});
const resolver = readGsdRuntimeProfileResolver(projectDir);
assert.ok(resolver !== null);
const entry = resolver.resolve('gsd-nonexistent-agent');
assert.strictEqual(entry, null, 'unknown agent name yields null');
});
});
describe('bug-2794: OpenCode agent install embeds model_profile_overrides model', () => {
let projectDir;
let homeDir;
let origHome;
let origCwd;
beforeEach(() => {
projectDir = makeTmp('proj');
homeDir = makeTmp('home');
origHome = process.env.HOME;
origCwd = process.cwd();
process.env.HOME = homeDir;
process.chdir(projectDir);
});
afterEach(() => {
if (origHome === undefined) delete process.env.HOME;
else process.env.HOME = origHome;
process.chdir(origCwd);
cleanup(projectDir);
cleanup(homeDir);
});
test('generated OpenCode agent frontmatter includes model from model_profile_overrides', () => {
writeJson(path.join(projectDir, '.planning', 'config.json'), {
runtime: 'opencode',
model_profile: 'balanced',
model_profile_overrides: {
opencode: {
sonnet: 'anthropic/claude-sonnet-4-7',
opus: 'anthropic/claude-opus-4-7',
haiku: 'anthropic/claude-haiku-4-5',
},
},
});
const oldLog = console.log;
console.log = () => {};
try {
install(false, 'opencode');
} finally {
console.log = oldLog;
}
const agentsDir = path.join(projectDir, '.opencode', 'agents');
assert.ok(fs.existsSync(agentsDir), 'agents directory should be created');
// gsd-roadmapper is balanced -> sonnet tier
const roadmapperPath = path.join(agentsDir, 'gsd-roadmapper.md');
assert.ok(fs.existsSync(roadmapperPath), 'gsd-roadmapper.md should exist');
const roadmapperContent = fs.readFileSync(roadmapperPath, 'utf-8');
assert.match(
roadmapperContent,
/^model: anthropic\/claude-sonnet-4-7$/m,
'gsd-roadmapper should have sonnet model from model_profile_overrides'
);
// gsd-planner is balanced -> opus tier
const plannerPath = path.join(agentsDir, 'gsd-planner.md');
assert.ok(fs.existsSync(plannerPath), 'gsd-planner.md should exist');
const plannerContent = fs.readFileSync(plannerPath, 'utf-8');
assert.match(
plannerContent,
/^model: anthropic\/claude-opus-4-7$/m,
'gsd-planner should have opus model from model_profile_overrides'
);
});
test('explicit model_overrides[agent] wins over model_profile_overrides tier', () => {
writeJson(path.join(projectDir, '.planning', 'config.json'), {
runtime: 'opencode',
model_profile: 'balanced',
model_overrides: {
'gsd-roadmapper': 'explicit-winner-model',
},
model_profile_overrides: {
opencode: {
sonnet: 'tier-model-that-should-lose',
},
},
});
const oldLog = console.log;
console.log = () => {};
try {
install(false, 'opencode');
} finally {
console.log = oldLog;
}
const roadmapperPath = path.join(projectDir, '.opencode', 'agents', 'gsd-roadmapper.md');
assert.ok(fs.existsSync(roadmapperPath));
const content = fs.readFileSync(roadmapperPath, 'utf-8');
assert.match(
content,
/^model: explicit-winner-model$/m,
'explicit model_overrides must win over model_profile_overrides tier'
);
assert.doesNotMatch(
content,
/tier-model-that-should-lose/,
'tier model must not appear when explicit override is present'
);
});
test('no model field when neither model_overrides nor model_profile_overrides is set', () => {
writeJson(path.join(projectDir, '.planning', 'config.json'), {
runtime: 'opencode',
model_profile: 'balanced',
});
const oldLog = console.log;
console.log = () => {};
try {
install(false, 'opencode');
} finally {
console.log = oldLog;
}
const roadmapperPath = path.join(projectDir, '.opencode', 'agents', 'gsd-roadmapper.md');
if (fs.existsSync(roadmapperPath)) {
const content = fs.readFileSync(roadmapperPath, 'utf-8');
// When no overrides, model field should either be absent or use built-in default
// The key invariant: no model field if there are no user-supplied overrides
// AND no built-in opencode defaults for this tier
// (gsd-roadmapper balanced = sonnet; opencode has built-in sonnet defaults)
// So we only assert no crash and no tier-model-not-provided entries
assert.ok(typeof content === 'string', 'agent file should be a string');
}
// Key: no exception thrown (test passes = no crash on missing overrides)
});
});