Files
msd-core/tests/claude-imperative-reference.test.cjs
Tom Boucher 69e7afd0c7 chore(#3212): bounded quantifiers over document content — prohibition with teeth — Phase 4 (#3441)
* feat(#3415): ship local/no-unbounded-quantifier, burn down ReDoS class

Phase 4 of epic #3212 (ADR-3212 §5/§7, the final phase). New rule flags
an unbounded */+/{n,} quantifier over a broad character class
([\s\S], dotAll ., or a 1-2-unit negated class like [^\n]/[^)\n] — the
exact #2128-fixed shape) applied to a regex whose match target is
data-flow-traced to readFileSync content.

eslint-rules/lib/readfilesync-trace.cjs extracts the data-flow tracer
shared with no-crlf-fragile-split (Phase 2) rather than a second copy
— no-crlf-fragile-split refactored onto it with zero behavior change,
parity-tested.

Real triage, not 798 mechanical edits: the ADR's census (2026-08-08)
screened every unbounded quantifier in the tree unscoped. Correctly
scoped to readFileSync-derived content (matching Phase 2's own G2/G3
scoping), the rule found 162 real hits across two detection waves — the
second wave (93) surfaced only after a genuine off-by-one bug in this
rule's own first draft was caught while writing its RuleTester tests
and fixed (the bug silently missed every directly-quantified [\s\S]*
with no gap before the quantifier — exactly the class this rule exists
to catch). 3 hits landed in production src/ (commands.cts, milestone.cts,
roadmap.cts) and were each empirically timed against adversarial input
(matching #2128's own measured-not-assumed precedent) — all confirmed
linear-time/benign, left unbounded with a measured-evidence comment
rather than mechanically bounded. The remaining 159 are test-file
fixture parsing (test-author-controlled, fixed-size content, not
adversarial input) — each suppressed with a specific, non-generic
reason. Zero functional behavior changed anywhere in this diff.

tests/no-pending-3212-markers.test.cjs locks the epic's own closing
invariant (ADR §7: "assert zero pending #3212 markers remain") — ground
truth confirmed trivially true today (no phase left any such marker
behind), now regression-locked going forward.

Design: .gsd/phase/chore-3415-prohibition-with-teeth/40-design.md
Test matrix: .gsd/phase/chore-3415-prohibition-with-teeth/50-test-matrix.md

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

* fix(#3415): correct rule category mislabel, add CI test-scope entry

An orthogonal Standards-axis review found eslint-rules/no-unbounded-quantifier.cjs
mistakenly carried meta.docs.category: 'Portability', copied from a sibling
rule without realizing what that implied: docs/contributing/cross-platform-
portability-rules.md governs an ADR-1703 rule family under a hard "zero
escape hatches" contract (tests/portability-rule-disable-ban.test.cjs's
PROTECTED_RULES bans eslint-disable for those rules entirely). This rule is
not part of that family — it's ADR-3212 (ReDoS/CWE-1333), a different epic —
and its eslint-disable-next-line suppressions (159 of them, added earlier
this same phase after empirical benign-verification) are an intentional,
correct design, not a bypass. Corrected to category: 'Best Practices',
matching the actual precedent (no-adhoc-regex-escape.cjs, Phase 1 of the
same epic, which is also correctly outside PROTECTED_RULES), and the rule's
own docstring now states this explicitly so a future reader doesn't have to
re-derive it.

Also registers a new scripts/ci-test-scope.cjs bucket so editing this rule
or the shared eslint-rules/lib/readfilesync-trace.cjs helper re-runs their
own test suites under targeted CI selection — was previously unregistered
and invisible to that fast-path (this PR's own gsd-test checkpoint runs the
full suite regardless, so this only affects future narrowly-scoped PRs).

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

* fix(#3415): bound no-unbounded-quantifier's own scanner (CWE-1333, ironic)

Security review found the rule meant to catch algorithmic-complexity bugs
had one of its own: hasUnboundedBroadQuantifier's negated-class inner
scan walked from each `[^` occurrence to the next `]` (or EOF) with no
bound, while the outer loop only ever advanced by one character — O(n²)
total work on a pattern with many unclosed `[^` runs. Runs unconditionally
inside checkPattern on any `new RegExp('literal string')` argument in any
linted file, before the (cheap) readFileSync data-flow gate — so a single
crafted string literal, no valid regex syntax required, could make
`npm run lint` / CI hang.

Empirically confirmed both the bug and the fix: pre-fix, n=4000/8000/
16000/32000 chars took 30.8/115.6/463.8/1874.3ms (~4x work per 2x n,
quadratic); extrapolated, the 300000-char repro from the finding would
run ~165s. Post-fix (bail the inner scan once units exceeds the rule's
own 1-2-unit scope, rather than continuing to hunt for a closing `]`),
the same 300000-char input runs in 8.7ms via the real rule module,
independently reconfirmed at 18ms via a fresh Linter.verify() call.

New regression row in tests/no-unbounded-quantifier.rule.test.cjs
asserts the RuleTester run on a 50000-char adversarial pattern
completes and returns a defined result — no wall-clock assertion
(CLAUDE.md Clock Seams / local/no-elapsed-assertion).

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

* fix(#3415): triage 3 new sites, re-raise ceiling after upstream batch

next merged 12 more PRs during this PR's review. Two consequences:

- tests/edit-phase.test.cjs (fix #3262, unrelated) added 3 new
  content.match(/<tag>([\s\S]*?)<\/tag>/) reads of this repo's own
  workflow .md content — the same Class A pattern as the ~159 sites
  already triaged elsewhere in this PR. Suppressed with the same
  established reason.
- lint-allow-test-rule-refs' ratchet ceiling needed re-raising again
  (301 -> 303) for the same reason as the two prior bumps: organic
  growth from unrelated, already-reviewed PRs landing concurrently,
  not a defect in this branch's own diff.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

---------

Co-authored-by: sim <sim@local>
Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com>
2026-08-14 10:02:28 -04:00

187 lines
10 KiB
JavaScript

// allow-test-rule: structural-regression-guard — AC2 requires asserting no `runtime === 'claude'` string-equality branch remains in bin/install.js — the descriptor-migration contract is a property of the source text, so a source-grep is the only faithful check (#2086)
'use strict';
/**
* claude imperative reference host — ADR-1239 Phase D / #2086 (EoS/claude).
*
* claude is GSD's tier-1 reference host — "golden parity vs. the Claude reference
* host" (ADR-1239 line 146). This proves claude is driven through the PUBLIC
* Host-Integration Interface (the imperative adapter), that its negotiated axes
* classify + negotiate correctly, that negotiation FAILS CLOSED on a corrupted
* descriptor, and that the migration retired the hardcoded `runtime === 'claude'`
* string-equality branches in bin/install.js (folded into descriptor-driven
* `runtime.hostBehaviors`).
*
* Mirrors tests/pi-imperative-reference.test.cjs + tests/vscode-ide-reference.test.cjs
* but binds against the REAL descriptor + the REAL installer source, since claude
* (unlike pi/vscode) has a real production install being folded in.
*/
const { test } = require('node:test');
const assert = require('node:assert/strict');
const fs = require('node:fs');
const path = require('node:path');
const { createImperativeAdapter } = require('../gsd-core/bin/lib/adapter-imperative.cjs');
const {
profileOf,
negotiateHostCapabilities,
PROTOCOL_VERSION,
PROFILE_BASELINES,
UNDOCUMENTED,
} = require('../gsd-core/bin/lib/host-integration.cjs');
const CLAUDE_CAP = JSON.parse(
fs.readFileSync(path.join(__dirname, '..', 'capabilities', 'claude', 'capability.json'), 'utf8'),
);
const CLAUDE_AXES = CLAUDE_CAP.runtime.hostIntegration;
// Requiring the installer (not as main) never runs the CLI; GSD_TEST_MODE is set
// defensively to match the install-test convention.
process.env.GSD_TEST_MODE = process.env.GSD_TEST_MODE || '1';
const installMod = require('../bin/install.js');
// -- AC2: driven through the public interface (imperative adapter) -----------
test('createImperativeAdapter classifies claude as imperative + composes the registry', () => {
const adapter = createImperativeAdapter({ runtime: 'claude' });
assert.equal(adapter.kind, 'imperative', "claude embeddingMode is imperative -> adapter kind must be 'imperative'");
assert.equal(adapter.runtime, 'claude');
assert.ok(adapter.registry && typeof adapter.registry === 'object', 'imperative adapter exposes the composed capability registry');
assert.equal(typeof adapter.install, 'function');
assert.equal(typeof adapter.uninstall, 'function');
});
test('claude descriptor embeddingMode agrees with the imperative adapter kind', () => {
assert.equal(CLAUDE_AXES.embeddingMode, 'imperative', 'capability.json must declare embeddingMode: imperative for the imperative binding');
});
test('claude axes classify as the programmatic-cli reference profile', () => {
assert.equal(profileOf(CLAUDE_AXES), 'programmatic-cli');
assert.notEqual(profileOf(CLAUDE_AXES), 'ide');
});
// -- AC3: every negotiated axis populated, negotiation is clean ---------------
test('claude negotiates its declared axes verbatim (no degradation of documented values)', () => {
const result = negotiateHostCapabilities({ ...CLAUDE_AXES, protocolVersion: PROTOCOL_VERSION });
assert.equal(result.protocolVersion, PROTOCOL_VERSION);
// Every declared scalar axis survives negotiation unchanged (all are known+documented).
assert.equal(result.effective.embeddingMode, 'imperative');
assert.equal(result.effective.commandSurface, CLAUDE_AXES.commandSurface);
assert.equal(result.effective.modelMode, CLAUDE_AXES.modelMode);
assert.equal(result.effective.hookBus, CLAUDE_AXES.hookBus);
assert.equal(result.effective.stateIO, CLAUDE_AXES.stateIO);
assert.equal(result.effective.transport, CLAUDE_AXES.transport);
assert.equal(result.effective.runtime, CLAUDE_AXES.runtime);
// No `undocumented` sentinel anywhere in claude's declared axes.
assert.ok(
!JSON.stringify(CLAUDE_AXES).includes(UNDOCUMENTED),
'claude descriptor must carry no `undocumented` sentinel (fully doc-sourced)',
);
});
// -- AC5: negotiation fails CLOSED on a corrupted / partial descriptor --------
test('negotiateHostCapabilities never throws for claude — even fully corrupted input', () => {
assert.doesNotThrow(() => negotiateHostCapabilities({}));
assert.doesNotThrow(() => negotiateHostCapabilities({ embeddingMode: UNDOCUMENTED }));
assert.doesNotThrow(() => negotiateHostCapabilities({ embeddingMode: 'wildly-unknown-future-value' }));
});
test('a partial/empty claude descriptor degrades to the safe floor — NOT the full programmatic-cli baseline', () => {
const result = negotiateHostCapabilities({});
// Fail-closed floor (SAFE_DEFAULTS), not the rich profile baseline.
assert.equal(result.effective.embeddingMode, 'declarative', 'omitted embeddingMode degrades closed to declarative');
assert.equal(result.effective.hookBus, 'none', 'omitted hookBus degrades closed to none');
assert.equal(result.effective.commandSurface, 'prose-only', 'omitted commandSurface degrades closed to prose-only');
assert.equal(result.effective.dispatch.namedDispatch, false, 'omitted dispatch degrades closed (no named dispatch)');
assert.notDeepEqual(
result.effective,
PROFILE_BASELINES['programmatic-cli'],
'a corrupted descriptor MUST NOT silently reuse the full programmatic-cli baseline',
);
assert.ok(result.warnings.length > 0, 'degrade-closed must surface warnings');
});
test('an undocumented/unknown claude axis value is not trusted (degraded closed per-axis)', () => {
const corrupted = { ...CLAUDE_AXES, embeddingMode: UNDOCUMENTED, hookBus: 'unknown-bus-kind' };
const result = negotiateHostCapabilities(corrupted);
assert.equal(result.effective.embeddingMode, 'declarative', 'undocumented embeddingMode -> safe floor');
assert.equal(result.effective.hookBus, 'none', 'unknown hookBus value -> safe floor');
// Untouched axes still negotiate to their declared (documented) values.
assert.equal(result.effective.stateIO, CLAUDE_AXES.stateIO);
});
// -- AC2: the hardcoded string-equality branches are retired ------------------
test('claude descriptor declares runtime.hostBehaviors (the folded-in host behaviors)', () => {
const hb = CLAUDE_CAP.runtime.hostBehaviors;
assert.ok(hb && typeof hb === 'object', 'capabilities/claude/capability.json must declare runtime.hostBehaviors');
// The behaviors that replaced the 13 `runtime === 'claude'` branches.
assert.equal(hb.permissionsSchema, 'claude');
assert.equal(hb.localInstallStyle, 'legacy-flat');
assert.equal(hb.sourceMarkerFile, '.gsd-source');
assert.equal(hb.authorsCanonicalWorkflow, true);
assert.equal(hb.ownsClaudePaths, true);
assert.equal(hb.nativeModelAliases, true);
assert.equal(hb.skillsGlobalOnboarding, true);
assert.equal(hb.attributionSource, 'settings-json-commit');
assert.deepEqual(hb.agentFrontmatterExtensions, ['effort']);
assert.equal(hb.settingsFileByScope.local, 'settings.local.json');
assert.equal(hb.settingsFileByScope.global, 'settings.json');
});
test('bin/install.js contains no `runtime === "claude"` / `runtime !== "claude"` string-equality branches (AC2)', () => {
const src = fs.readFileSync(path.join(__dirname, '..', 'bin', 'install.js'), 'utf8');
// Strip comments + backtick/inline-code spans so PROSE mentions of the old
// pattern (a comment explaining "not a string-equality branch") do not
// false-positive — only LIVE code counts.
const codeOnly = src
// eslint-disable-next-line local/no-unbounded-quantifier -- parses this repo's own bounded bin/install.js source, not adversarial input
.replace(/\/\*[\s\S]*?\*\//g, '') // block comments
// eslint-disable-next-line local/no-unbounded-quantifier -- parses this repo's own bounded bin/install.js source, not adversarial input
.replace(/\/\/[^\r\n]*/g, '') // line comments (CRLF-safe)
// eslint-disable-next-line local/no-unbounded-quantifier -- parses this repo's own bounded bin/install.js source, not adversarial input
.replace(/`[^`]*`/g, ''); // backtick / inline-code spans
const offenders = codeOnly.match(/runtime\s*[!=]==\s*'claude'/g) || [];
assert.deepEqual(
offenders,
[],
`AC2: every hardcoded runtime==='claude'/!=='claude' branch must be descriptor-driven; found: ${offenders.join(', ')}`,
);
});
// -- Reviewer #2106 (elevated): #338 privacy fail-safe on registry-load failure --
// If the first-party capability registry fails to load, `_hostBehaviors('claude')`
// would return {} and route a claude LOCAL install to the repo-shared settings.json
// instead of the gitignored settings.local.json — silently reintroducing #338. The
// reference host must degrade CLOSED (safe) for its privacy-critical keys.
test('claude #338-critical host behaviors degrade CLOSED when the capability registry cannot load', () => {
// Simulate a broken bundle: registry is undefined.
const degraded = installMod._resolveHostBehaviors('claude', undefined);
assert.equal(degraded.settingsFileByScope.local, 'settings.local.json',
'#338: a claude LOCAL install must still route to the gitignored settings.local.json');
assert.equal(degraded.settingsFileByScope.global, 'settings.json');
assert.equal(degraded.permissionsSchema, 'claude', 'permission cleanup/merge must still apply');
assert.equal(degraded.sourceMarkerFile, '.gsd-source');
});
test('with the registry present, claude host behaviors come from the live descriptor (superset of the fail-safe floor)', () => {
const reg = require('../gsd-core/bin/lib/capability-registry.cjs');
const declared = installMod._resolveHostBehaviors('claude', reg);
assert.equal(declared.settingsFileByScope.local, 'settings.local.json');
assert.equal(declared.localInstallStyle, 'legacy-flat');
assert.equal(declared.authorsCanonicalWorkflow, true);
// The fail-safe floor is a strict subset of what the descriptor declares.
for (const k of Object.keys(installMod.FALLBACK_HOST_BEHAVIORS.claude)) {
assert.ok(k in declared, `descriptor must still declare the #338-critical key '${k}'`);
}
});
test('a non-reference runtime has no fail-safe fallback (degrades to the generic path)', () => {
assert.deepEqual(installMod._resolveHostBehaviors('opencode', undefined), {});
assert.deepEqual(installMod._resolveHostBehaviors('codex', undefined), {});
});