sneakoscope 4.1.0 → 4.2.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +16 -3
- package/crates/sks-core/Cargo.lock +1 -1
- package/crates/sks-core/Cargo.toml +1 -1
- package/crates/sks-core/src/main.rs +1 -1
- package/dist/bin/sks.js +1 -1
- package/dist/cli/command-registry.js +1 -1
- package/dist/cli/router.js +6 -1
- package/dist/commands/doctor.js +272 -127
- package/dist/core/auto-review.js +1 -1
- package/dist/core/codex/agent-config-file-repair.js +43 -2
- package/dist/core/codex-app/codex-agent-role-sync.js +4 -4
- package/dist/core/codex-control/codex-0142-capability.js +51 -6
- package/dist/core/codex-control/codex-app-server-v2-client.js +2 -2
- package/dist/core/codex-native/codex-native-feature-broker.js +50 -0
- package/dist/core/codex-native/native-capability-postcheck.js +59 -16
- package/dist/core/codex-native/native-capability-repair-matrix.js +77 -13
- package/dist/core/commands/mad-db-command.js +146 -51
- package/dist/core/commands/mad-sks-command.js +51 -61
- package/dist/core/db-safety.js +35 -37
- package/dist/core/doctor/doctor-dirty-planner.js +9 -4
- package/dist/core/doctor/doctor-native-capability-repair.js +42 -7
- package/dist/core/doctor/doctor-readiness-matrix.js +9 -5
- package/dist/core/doctor/doctor-repair-postcheck.js +10 -1
- package/dist/core/doctor/doctor-transaction.js +1 -1
- package/dist/core/doctor/supabase-mcp-repair.js +2 -2
- package/dist/core/feature-registry.js +1 -1
- package/dist/core/fsx.js +1 -1
- package/dist/core/init.js +5 -4
- package/dist/core/mad-db/mad-db-capability.js +203 -74
- package/dist/core/mad-db/mad-db-coordinator.js +287 -0
- package/dist/core/mad-db/mad-db-executor.js +156 -0
- package/dist/core/mad-db/mad-db-ledger.js +1 -1
- package/dist/core/mad-db/mad-db-lock.js +40 -0
- package/dist/core/mad-db/mad-db-operation-store.js +140 -0
- package/dist/core/mad-db/mad-db-policy-resolver.js +42 -22
- package/dist/core/mad-db/mad-db-policy.js +195 -0
- package/dist/core/mad-db/mad-db-postconditions.js +30 -0
- package/dist/core/mad-db/mad-db-recovery.js +27 -0
- package/dist/core/mad-db/mad-db-result-lifecycle.js +31 -102
- package/dist/core/mad-db/mad-db-runtime-profile.js +121 -0
- package/dist/core/mad-db/mad-db-target.js +64 -0
- package/dist/core/managed-assets/managed-assets-manifest.js +14 -4
- package/dist/core/pipeline-internals/runtime-core.js +40 -0
- package/dist/core/providers/glm/bench/glm-benchmark-runner.js +4 -3
- package/dist/core/providers/glm/bench/glm-benchmark-types.js +1 -1
- package/dist/core/release/release-gate-dag.js +6 -5
- package/dist/core/routes.js +23 -8
- package/dist/core/update/update-migration-state.js +265 -50
- package/dist/core/update-check.js +6 -6
- package/dist/core/version.js +1 -1
- package/dist/core/zellij/zellij-launcher.js +17 -5
- package/dist/core/zellij/zellij-slot-column-anchor.js +5 -1
- package/dist/scripts/check-dist-runtime.js +3 -2
- package/dist/scripts/codex-0142-manifest-check.js +2 -1
- package/dist/scripts/config-managed-merge-callsite-coverage-check.js +6 -0
- package/dist/scripts/doctor-dirty-plan-check.js +1 -1
- package/dist/scripts/doctor-transaction-engine-check.js +1 -0
- package/dist/scripts/doctor-warning-only-not-blocker-check.js +18 -1
- package/dist/scripts/loop-directive-check-lib.js +2 -1
- package/dist/scripts/mad-db-capability-check.js +13 -2
- package/dist/scripts/mad-db-command-check.js +7 -5
- package/dist/scripts/mad-db-hook-idempotency-check.js +21 -0
- package/dist/scripts/mad-db-ledger-check.js +2 -1
- package/dist/scripts/mad-db-lifecycle-hook-decision-check.js +5 -4
- package/dist/scripts/mad-db-mad-command-check.js +29 -16
- package/dist/scripts/mad-db-mcp-result-lifecycle-check.js +11 -10
- package/dist/scripts/mad-db-one-cycle-bounded-check.js +15 -18
- package/dist/scripts/mad-db-one-cycle-consumption-check.js +3 -3
- package/dist/scripts/mad-db-operation-lifecycle-blackbox.js +9 -9
- package/dist/scripts/mad-db-operation-lifecycle-ledger-check.js +6 -6
- package/dist/scripts/mad-db-parallel-lifecycle-check.js +24 -0
- package/dist/scripts/mad-db-policy-v2-check.js +20 -0
- package/dist/scripts/mad-db-priority-resolver-check.js +5 -5
- package/dist/scripts/mad-db-real-supabase-e2e.js +166 -0
- package/dist/scripts/mad-db-route-identity-check.js +28 -0
- package/dist/scripts/mad-db-runtime-profile-lifecycle-check.js +24 -0
- package/dist/scripts/mad-db-safety-conflict-matrix-check.js +3 -3
- package/dist/scripts/mad-db-skill-policy-snapshot-check.js +15 -0
- package/dist/scripts/mad-sks-zellij-launch-check.js +7 -1
- package/dist/scripts/managed-role-manifest-parity-check.js +4 -1
- package/dist/scripts/naruto-real-parallelism-blackbox.js +17 -4
- package/dist/scripts/native-capability-postcheck-check.js +1 -0
- package/dist/scripts/native-capability-repair-matrix-check.js +2 -0
- package/dist/scripts/native-chrome-web-review-repair-check.js +1 -0
- package/dist/scripts/native-computer-use-repair-check.js +1 -0
- package/dist/scripts/release-dag-full-coverage-check.js +6 -0
- package/dist/scripts/release-triwiki-first-runner-blackbox.js +5 -1
- package/dist/scripts/sks-3-1-5-directive-check-lib.js +1 -1
- package/dist/scripts/sks-401-all-feature-regression-blackbox.js +1 -1
- package/dist/scripts/update-concurrent-lock-check.js +1 -0
- package/dist/scripts/update-first-command-migration-check.js +4 -3
- package/package.json +13 -2
- package/schemas/mad-db/mad-db-capability.schema.json +92 -19
- package/schemas/update-migration.schema.json +13 -1
|
@@ -0,0 +1,64 @@
|
|
|
1
|
+
import path from 'node:path';
|
|
2
|
+
import { exists, readText, sha256 } from '../fsx.js';
|
|
3
|
+
export async function resolveMadDbTarget(root, input = {}) {
|
|
4
|
+
const args = input.args || [];
|
|
5
|
+
const explicit = input.projectRef || readOption(args, '--project-ref', '') || process.env.SKS_MAD_DB_PROJECT_REF || process.env.SKS_MAD_DB_E2E_PROJECT_REF || '';
|
|
6
|
+
const candidates = explicit ? [explicit] : await projectRefCandidates(root);
|
|
7
|
+
const projectRef = explicit || (candidates.length === 1 ? candidates[0] || '' : '');
|
|
8
|
+
const target = normalizeTarget(input.target || readOption(args, '--target', '') || process.env.SKS_MAD_DB_TARGET || process.env.SKS_MAD_DB_E2E_TARGET || 'production');
|
|
9
|
+
const allowedSchemas = input.allowedSchemas?.length
|
|
10
|
+
? input.allowedSchemas
|
|
11
|
+
: splitCsv(readOption(args, '--schema', readOption(args, '--schemas', process.env.SKS_MAD_DB_SCHEMAS || 'public')));
|
|
12
|
+
const blockers = [];
|
|
13
|
+
if (!projectRef)
|
|
14
|
+
blockers.push(candidates.length > 1 ? 'mad_db_project_ref_ambiguous' : 'mad_db_project_ref_missing');
|
|
15
|
+
return {
|
|
16
|
+
schema: 'sks.mad-db-target.v1',
|
|
17
|
+
project_ref: projectRef || null,
|
|
18
|
+
project_ref_hash: projectRef ? sha256(projectRef).slice(0, 16) : null,
|
|
19
|
+
target_environment: target,
|
|
20
|
+
allowed_schemas: allowedSchemas.length ? allowedSchemas : ['public'],
|
|
21
|
+
source: explicit ? 'explicit_or_environment' : candidates.length === 1 ? 'managed_config_single_candidate' : 'unresolved',
|
|
22
|
+
blockers,
|
|
23
|
+
candidates: candidates.map((candidate) => `${sha256(candidate).slice(0, 8)}:${candidate.slice(0, 2)}...`)
|
|
24
|
+
};
|
|
25
|
+
}
|
|
26
|
+
export async function projectRootHash(root) {
|
|
27
|
+
return sha256(path.resolve(root)).slice(0, 24);
|
|
28
|
+
}
|
|
29
|
+
async function projectRefCandidates(root) {
|
|
30
|
+
const files = ['.codex/config.toml', '.mcp.json', 'mcp.json', '.cursor/mcp.json', '.vscode/mcp.json'];
|
|
31
|
+
const out = new Set();
|
|
32
|
+
for (const rel of files) {
|
|
33
|
+
const file = path.join(root, rel);
|
|
34
|
+
if (!(await exists(file)))
|
|
35
|
+
continue;
|
|
36
|
+
const text = await readText(file, '');
|
|
37
|
+
for (const match of String(text).matchAll(/project_ref=([a-z0-9_-]+)/gi))
|
|
38
|
+
out.add(match[1] || '');
|
|
39
|
+
for (const match of String(text).matchAll(/project_ref["']?\s*[:=]\s*["']([a-z0-9_-]+)["']/gi))
|
|
40
|
+
out.add(match[1] || '');
|
|
41
|
+
}
|
|
42
|
+
return [...out].filter(Boolean);
|
|
43
|
+
}
|
|
44
|
+
function normalizeTarget(value) {
|
|
45
|
+
const text = String(value || '').toLowerCase();
|
|
46
|
+
if (text === 'local')
|
|
47
|
+
return 'local';
|
|
48
|
+
if (text === 'branch')
|
|
49
|
+
return 'branch';
|
|
50
|
+
if (text === 'preview')
|
|
51
|
+
return 'preview';
|
|
52
|
+
return 'production';
|
|
53
|
+
}
|
|
54
|
+
function splitCsv(value) {
|
|
55
|
+
return String(value || '').split(',').map((entry) => entry.trim()).filter(Boolean);
|
|
56
|
+
}
|
|
57
|
+
function readOption(args, name, fallback) {
|
|
58
|
+
const index = args.indexOf(name);
|
|
59
|
+
if (index >= 0 && args[index + 1] && !String(args[index + 1]).startsWith('--'))
|
|
60
|
+
return String(args[index + 1]);
|
|
61
|
+
const prefixed = args.find((arg) => String(arg).startsWith(`${name}=`));
|
|
62
|
+
return prefixed ? prefixed.slice(name.length + 1) : fallback;
|
|
63
|
+
}
|
|
64
|
+
//# sourceMappingURL=mad-db-target.js.map
|
|
@@ -1,15 +1,15 @@
|
|
|
1
1
|
import { REQUIRED_CODEX_MODEL } from '../codex-model-guard.js';
|
|
2
2
|
export const MANAGED_ASSET_SCHEMA_VERSION = 1;
|
|
3
|
-
export const MANAGED_ASSET_VERSION = '4.
|
|
3
|
+
export const MANAGED_ASSET_VERSION = '4.2.0';
|
|
4
4
|
export const MANAGED_ASSET_MARKER = 'SKS-MANAGED-ASSET';
|
|
5
5
|
export const MANAGED_AGENT_ROLES = Object.freeze([
|
|
6
6
|
role('sks-explorer', 'analysis-scout.toml', 'analysis_scout', 'Read-only SKS analysis scout retained for stale Codex agent-role config repair.', 'read-only', ['analysis-scout', 'analysis_scout', 'native_agent']),
|
|
7
7
|
role('sks-planner', 'team-consensus.toml', 'team_consensus', 'Planning and debate specialist for SKS Team mode.', 'read-only', ['team-consensus', 'team_consensus']),
|
|
8
8
|
role('sks-implementer', 'implementation-worker.toml', 'implementation_worker', 'Implementation specialist for bounded SKS Team write sets.', 'workspace-write', ['implementation-worker', 'implementation_worker']),
|
|
9
9
|
role('sks-checker', 'qa-reviewer.toml', 'qa_reviewer', 'Strict read-only verification reviewer for correctness, regressions, and final evidence.', 'read-only', ['qa-reviewer', 'qa_reviewer']),
|
|
10
|
-
role('sks-release-verifier', '
|
|
11
|
-
role('sks-zellij-ui-verifier', '
|
|
12
|
-
role('sks-codex-probe-verifier', '
|
|
10
|
+
role('sks-release-verifier', 'sks-release-verifier.toml', 'sks_release_verifier', 'Read-only release verifier for repository, docs, tests, API, and risk slices.', 'read-only', ['native-agent-intake', 'native_agent', 'release-verifier']),
|
|
11
|
+
role('sks-zellij-ui-verifier', 'sks-zellij-ui-verifier.toml', 'sks_zellij_ui_verifier', 'Read-only Zellij UI verifier for session, pane, layout, and terminal evidence.', 'read-only', ['native-agent-intake', 'native_agent', 'zellij-ui-verifier']),
|
|
12
|
+
role('sks-codex-probe-verifier', 'sks-codex-probe-verifier.toml', 'sks_codex_probe_verifier', 'Read-only Codex probe verifier for CLI, App, SDK, MCP, and native capability evidence.', 'read-only', ['native-agent-intake', 'native_agent', 'codex-probe-verifier']),
|
|
13
13
|
role('db-safety-reviewer', 'db-safety-reviewer.toml', 'db_safety_reviewer', 'Read-only database safety reviewer for SQL, migrations, Supabase, and rollback safety.', 'read-only', ['db-safety-reviewer', 'db_safety_reviewer'])
|
|
14
14
|
]);
|
|
15
15
|
export const MANAGED_SKILLS = Object.freeze([
|
|
@@ -41,6 +41,7 @@ export const CONTEXT7_MANAGED_SERVER = Object.freeze({
|
|
|
41
41
|
});
|
|
42
42
|
export function managedAgentRoleByFile(filename) {
|
|
43
43
|
const base = filename.split(/[\\/]/).pop() || filename;
|
|
44
|
+
assertUniqueManagedAgentRoleFilenames();
|
|
44
45
|
return MANAGED_AGENT_ROLES.find((role) => role.filename === base) || null;
|
|
45
46
|
}
|
|
46
47
|
export function managedAgentRoleByName(name) {
|
|
@@ -87,6 +88,15 @@ export function managedAgentRoleOwnsText(text, role) {
|
|
|
87
88
|
export function normalizeRoleName(name) {
|
|
88
89
|
return String(name || '').trim().replace(/\.toml$/i, '').replace(/_/g, '-').toLowerCase();
|
|
89
90
|
}
|
|
91
|
+
export function assertUniqueManagedAgentRoleFilenames() {
|
|
92
|
+
const seen = new Map();
|
|
93
|
+
for (const role of MANAGED_AGENT_ROLES) {
|
|
94
|
+
const existing = seen.get(role.filename);
|
|
95
|
+
if (existing)
|
|
96
|
+
throw new Error(`duplicate managed agent role filename: ${role.filename} for ${existing} and ${role.id}`);
|
|
97
|
+
seen.set(role.filename, role.id);
|
|
98
|
+
}
|
|
99
|
+
}
|
|
90
100
|
function role(id, filename, codexName, description, sandbox, aliases) {
|
|
91
101
|
return {
|
|
92
102
|
id,
|
|
@@ -22,6 +22,7 @@ import { SPEED_LANE_POLICY } from '../proof-field.js';
|
|
|
22
22
|
import { validateRouteCompletionProof } from '../proof/route-proof-gate.js';
|
|
23
23
|
import { routeFromState, routeRequiresCompletionProof } from '../proof/route-proof-policy.js';
|
|
24
24
|
import { permissionGateSummary } from '../permission-gates.js';
|
|
25
|
+
import { prepareMadDbMission } from '../mad-db/mad-db-coordinator.js';
|
|
25
26
|
import { AGENT_INTAKE_STAGE_ID, AGENT_COUNT } from '../agents/agent-schema.js';
|
|
26
27
|
import { normalizeAgentPolicy, routeRequiresAgentIntake, agentPipelineStage } from '../agents/agent-plan.js';
|
|
27
28
|
import { readAgentGateStatus } from '../agents/agent-gate.js';
|
|
@@ -458,6 +459,8 @@ export function promptPipelineContext(prompt, route = null) {
|
|
|
458
459
|
lines.push('AutoResearch route: load autoresearch-loop plus seo-geo-optimizer when SEO/GEO, discoverability, README, npm, GitHub stars, ranking, or AI-search visibility is relevant.');
|
|
459
460
|
if (route?.id === 'DB')
|
|
460
461
|
lines.push('DB route: scan/check database risk first; destructive DB operations remain forbidden.');
|
|
462
|
+
if (route?.id === 'MadDB')
|
|
463
|
+
lines.push('MadDB route: explicit invocation is the SQL-plane approval boundary. Use the mission-local write-capable Supabase MCP profile only for the bound cycle, verify execute_sql/apply_migration inventory before claiming ready, execute requested SQL-plane mutations, read back postconditions, then close the capability/profile and prove normal read-only restoration. Supabase project/account/billing/credential control-plane actions remain denied.');
|
|
461
464
|
if (route?.id === 'GX')
|
|
462
465
|
lines.push('GX route: use deterministic vgraph/beta render, validate, drift, and snapshot artifacts.');
|
|
463
466
|
return lines.join('\n');
|
|
@@ -517,6 +520,8 @@ export async function prepareRoute(root, prompt, state = {}) {
|
|
|
517
520
|
return withSkillDreamContext(await prepareGoal(root, route, task, routeNeedsContext7(route, cleanPrompt)), dreamContext);
|
|
518
521
|
if (route.id === 'ImageUXReview')
|
|
519
522
|
return withSkillDreamContext(await prepareImageUxReview(root, route, task, routeNeedsContext7(route, cleanPrompt)), dreamContext);
|
|
523
|
+
if (route.id === 'MadDB')
|
|
524
|
+
return withSkillDreamContext(await prepareMadDb(root, route, task, routeNeedsContext7(route, cleanPrompt)), dreamContext);
|
|
520
525
|
const required = routeNeedsContext7(route, cleanPrompt);
|
|
521
526
|
const reasoning = routeReasoning(route, cleanPrompt);
|
|
522
527
|
const nativeSessionsRequired = routeRequiresSubagents(route, cleanPrompt);
|
|
@@ -1111,6 +1116,41 @@ async function prepareDb(root, route, task, required) {
|
|
|
1111
1116
|
await setCurrent(root, routeState(id, route, 'DB_REVIEW_REQUIRED', required, { prompt: task, pipeline_plan_ready: validatePipelinePlan(pipelinePlan).ok, pipeline_plan_path: PIPELINE_PLAN_ARTIFACT }));
|
|
1112
1117
|
return routeContext(route, id, task, required, 'Run sks db policy/scan/check as needed, keep DB operations read-only, record safe MCP policy, and pass db-review.json.');
|
|
1113
1118
|
}
|
|
1119
|
+
async function prepareMadDb(root, route, task, required) {
|
|
1120
|
+
const prepared = await prepareMadDbMission({ root, task, verifyTools: false });
|
|
1121
|
+
const dir = missionDir(root, prepared.mission_id);
|
|
1122
|
+
const pipelinePlan = await writePipelinePlan(dir, {
|
|
1123
|
+
missionId: prepared.mission_id,
|
|
1124
|
+
route,
|
|
1125
|
+
task,
|
|
1126
|
+
required,
|
|
1127
|
+
ambiguity: { required: true, slots: 0, auto_sealed: true, passed: true, contract_hash: prepared.capability.operator_intent_hash }
|
|
1128
|
+
});
|
|
1129
|
+
await appendJsonl(path.join(dir, 'events.jsonl'), {
|
|
1130
|
+
ts: nowIso(),
|
|
1131
|
+
type: 'mad_db.route_prepared',
|
|
1132
|
+
mission_id: prepared.mission_id,
|
|
1133
|
+
cycle_id: prepared.cycle_id,
|
|
1134
|
+
blockers: prepared.blockers
|
|
1135
|
+
});
|
|
1136
|
+
await setCurrent(root, routeState(prepared.mission_id, route, prepared.ok ? 'MADDB_SQL_PLANE_CAPABILITY_ACTIVE' : 'MADDB_BLOCKED', required, {
|
|
1137
|
+
prompt: task,
|
|
1138
|
+
questions_allowed: false,
|
|
1139
|
+
implementation_allowed: prepared.ok,
|
|
1140
|
+
ambiguity_gate_required: true,
|
|
1141
|
+
ambiguity_gate_passed: true,
|
|
1142
|
+
mad_db_active: prepared.ok,
|
|
1143
|
+
mad_db_cycle_id: prepared.cycle_id,
|
|
1144
|
+
mad_db_runtime_session_id: prepared.capability.runtime_session_id,
|
|
1145
|
+
mad_db_profile_sha256: prepared.capability.transport.profile_sha256,
|
|
1146
|
+
mad_db_capability_mission_id: prepared.mission_id,
|
|
1147
|
+
mad_db_capability_file: 'mad-db-capability.json',
|
|
1148
|
+
stop_gate: 'mad-db-gate.json',
|
|
1149
|
+
pipeline_plan_ready: validatePipelinePlan(pipelinePlan).ok,
|
|
1150
|
+
pipeline_plan_path: PIPELINE_PLAN_ARTIFACT
|
|
1151
|
+
}));
|
|
1152
|
+
return routeContext(route, prepared.mission_id, task, required, `MadDB mission/capability/profile were created atomically for cycle ${prepared.cycle_id}. Verify Supabase MCP tool inventory exposes execute_sql and apply_migration before claiming ready; after execution require read-back proof, finally close the profile/capability and prove read-only restoration.`);
|
|
1153
|
+
}
|
|
1114
1154
|
async function prepareGx(root, route, task, required) {
|
|
1115
1155
|
const { id, dir } = await createMission(root, { mode: 'gx', prompt: task });
|
|
1116
1156
|
await writeJsonAtomic(path.join(dir, 'gx-gate.json'), { passed: false, vgraph_beta_render: false, validation: false, drift_snapshot: false, context7_evidence: false });
|
|
@@ -13,6 +13,7 @@ import { runGlmDirectBenchCase } from './glm-direct-bench-runner.js';
|
|
|
13
13
|
import { computeGlmBenchmarkComparison } from './glm-bench-comparison.js';
|
|
14
14
|
import { buildGlmBenchModelLockProof } from './glm-bench-model-lock-proof.js';
|
|
15
15
|
import { writeGlmBenchReport } from './glm-bench-report.js';
|
|
16
|
+
import { GLM_BENCHMARK_VERSION } from './glm-benchmark-types.js';
|
|
16
17
|
const NARUTO_WORKER_COUNTS = [1, 4, 8, 12];
|
|
17
18
|
export async function runGlmBenchmark(root, args = [], deps = {}) {
|
|
18
19
|
const live = args.includes('--live');
|
|
@@ -130,7 +131,7 @@ export async function runGlmBenchmark(root, args = [], deps = {}) {
|
|
|
130
131
|
};
|
|
131
132
|
const result = {
|
|
132
133
|
schema: 'sks.glm-benchmark-result.v1',
|
|
133
|
-
version:
|
|
134
|
+
version: GLM_BENCHMARK_VERSION,
|
|
134
135
|
generated_at: nowIso(),
|
|
135
136
|
status: 'live',
|
|
136
137
|
model: GLM_52_OPENROUTER_MODEL,
|
|
@@ -157,7 +158,7 @@ export async function runGlmBenchmark(root, args = [], deps = {}) {
|
|
|
157
158
|
function dryRunResult(root, startedMs) {
|
|
158
159
|
return {
|
|
159
160
|
schema: 'sks.glm-benchmark-result.v1',
|
|
160
|
-
version:
|
|
161
|
+
version: GLM_BENCHMARK_VERSION,
|
|
161
162
|
generated_at: nowIso(),
|
|
162
163
|
status: 'dry_run',
|
|
163
164
|
model: GLM_52_OPENROUTER_MODEL,
|
|
@@ -180,7 +181,7 @@ function dryRunResult(root, startedMs) {
|
|
|
180
181
|
function blockedResult(root, warnings) {
|
|
181
182
|
return {
|
|
182
183
|
schema: 'sks.glm-benchmark-result.v1',
|
|
183
|
-
version:
|
|
184
|
+
version: GLM_BENCHMARK_VERSION,
|
|
184
185
|
generated_at: nowIso(),
|
|
185
186
|
status: 'blocked',
|
|
186
187
|
model: GLM_52_OPENROUTER_MODEL,
|
|
@@ -1,2 +1,2 @@
|
|
|
1
|
-
export const GLM_BENCHMARK_VERSION = '4.
|
|
1
|
+
export const GLM_BENCHMARK_VERSION = '4.2.0';
|
|
2
2
|
//# sourceMappingURL=glm-benchmark-types.js.map
|
|
@@ -28,15 +28,16 @@ export async function runReleaseGateDag(input) {
|
|
|
28
28
|
const preset = input.preset || 'release';
|
|
29
29
|
const manifest = loadReleaseGateManifest(root);
|
|
30
30
|
const presetGates = selectReleaseGatePreset(manifest, preset);
|
|
31
|
-
const triwikiGraph = input.triwiki !== false && (preset === 'affected' || preset === 'fast' || preset === 'confidence') && input.full !== true
|
|
32
|
-
? computeTriWikiAffectedGraph({ root, tier: preset === 'fast' ? 'affected' : 'confidence', changedSince: input.changedSince || 'auto', ...(input.changedFiles ? { changedFiles: input.changedFiles } : {}) })
|
|
33
|
-
: null;
|
|
34
31
|
const affected = (preset === 'affected' || preset === 'fast' || preset === 'confidence') && input.full !== true
|
|
35
32
|
? selectAffectedReleaseGates(root, manifest, presetGates, { changedSince: input.changedSince || 'auto', ...(input.changedFiles ? { changedFiles: input.changedFiles } : {}), preset })
|
|
36
33
|
: selectAffectedReleaseGates(root, manifest, presetGates, { full: true, preset });
|
|
34
|
+
const rootReleaseSurfaceChanged = affected.selection.changed_files.some((file) => file === 'package.json' || file === 'package-lock.json' || file === 'release-gates.v2.json');
|
|
35
|
+
const triwikiGraph = input.triwiki !== false && !rootReleaseSurfaceChanged && (preset === 'affected' || preset === 'fast' || preset === 'confidence') && input.full !== true
|
|
36
|
+
? computeTriWikiAffectedGraph({ root, tier: preset === 'fast' ? 'affected' : 'confidence', changedSince: input.changedSince || 'auto', ...(input.changedFiles ? { changedFiles: input.changedFiles } : {}) })
|
|
37
|
+
: null;
|
|
37
38
|
const triwikiSelectionUsed = Boolean(triwikiGraph);
|
|
38
39
|
const triwikiConservative = Boolean(triwikiGraph?.conservative_reason);
|
|
39
|
-
const triwikiSelectedIds = new Set(triwikiGraph && !triwikiConservative ? triwikiGraph.gates :
|
|
40
|
+
const triwikiSelectedIds = new Set(triwikiGraph && !triwikiConservative ? triwikiGraph.gates : affected.gates.map((gate) => gate.id));
|
|
40
41
|
const selected = triwikiGraph
|
|
41
42
|
? presetGates.filter((gate) => triwikiSelectedIds.has(gate.id))
|
|
42
43
|
: affected.gates;
|
|
@@ -45,7 +46,7 @@ export async function runReleaseGateDag(input) {
|
|
|
45
46
|
affected.selection.mode = 'affected';
|
|
46
47
|
affected.selection.selected_gate_ids = selected.map((gate) => gate.id);
|
|
47
48
|
affected.selection.skipped_gate_ids = triwikiSkippedGates;
|
|
48
|
-
affected.selection.reasons = Object.fromEntries(selected.map((gate) => [gate.id, triwikiConservative ? `
|
|
49
|
+
affected.selection.reasons = Object.fromEntries(selected.map((gate) => [gate.id, triwikiConservative ? `triwiki_conservative_fallback:${triwikiGraph.conservative_reason}` : 'triwiki-affected']));
|
|
49
50
|
}
|
|
50
51
|
const selectedIds = new Set(selected.map((gate) => gate.id));
|
|
51
52
|
const affectedExternalSatisfiedDeps = affected.selection.mode === 'affected'
|
package/dist/core/routes.js
CHANGED
|
@@ -231,7 +231,7 @@ export function stackCurrentDocsPolicy(commandPrefix = 'sks') {
|
|
|
231
231
|
validate_command: `${prefix} wiki validate .sneakoscope/wiki/context-pack.json`,
|
|
232
232
|
priority: 'must_precede_coding_style_defaults',
|
|
233
233
|
examples: [
|
|
234
|
-
'Supabase hosted projects should prefer sb_publishable_ and sb_secret_ keys over legacy anon
|
|
234
|
+
'Supabase hosted projects should prefer sb_publishable_ and sb_secret_ keys over legacy anon and service role keys when current docs apply.',
|
|
235
235
|
'Next.js 16 deprecates the middleware file convention in favor of proxy.ts/proxy.js.',
|
|
236
236
|
'Vercel Function duration limits, including the 300s default with Fluid Compute, are deployment constraints that must shape long-running server work.'
|
|
237
237
|
]
|
|
@@ -239,7 +239,7 @@ export function stackCurrentDocsPolicy(commandPrefix = 'sks') {
|
|
|
239
239
|
}
|
|
240
240
|
export function stackCurrentDocsPolicyText(commandPrefix = 'sks') {
|
|
241
241
|
const policy = stackCurrentDocsPolicy(commandPrefix);
|
|
242
|
-
return `Stack current-docs policy: whenever project tech stack is added or a framework/package/runtime/platform version changes, fetch current docs with Context7 (resolve-library-id then query-docs) or official vendor web docs before coding, record the syntax/limits/security guidance as high-priority TriWiki claims in ${policy.memory_path}, run "${policy.refresh_command}", then "${policy.validate_command}". Treat these claims as higher priority than model-memory defaults. Examples include Supabase publishable/secret keys replacing legacy anon
|
|
242
|
+
return `Stack current-docs policy: whenever project tech stack is added or a framework/package/runtime/platform version changes, fetch current docs with Context7 (resolve-library-id then query-docs) or official vendor web docs before coding, record the syntax/limits/security guidance as high-priority TriWiki claims in ${policy.memory_path}, run "${policy.refresh_command}", then "${policy.validate_command}". Treat these claims as higher priority than model-memory defaults. Examples include Supabase publishable/secret keys replacing legacy anon and service role guidance for hosted projects, Next.js 16 proxy.ts/proxy.js replacing the deprecated middleware file convention, avoiding stale webpack defaults when newer framework guidance says otherwise, and Vercel Function duration limits such as the 300s default under Fluid Compute.`;
|
|
243
243
|
}
|
|
244
244
|
export function triwikiContextTrackingText(commandPrefix = 'sks') {
|
|
245
245
|
const ctx = triwikiContextTracking(commandPrefix);
|
|
@@ -549,21 +549,34 @@ export const ROUTES = [
|
|
|
549
549
|
cliEntrypoint: 'sks db scan',
|
|
550
550
|
examples: ['$DB check this migration safely']
|
|
551
551
|
},
|
|
552
|
+
{
|
|
553
|
+
id: 'MadDB',
|
|
554
|
+
command: '$MAD-DB',
|
|
555
|
+
mode: 'MADDB',
|
|
556
|
+
route: 'first-class MadDB SQL-plane execution',
|
|
557
|
+
description: 'Explicit one-cycle MadDB route. When invoked by $MAD-DB or sks mad-db run|exec|apply-migration, SQL-plane mutations such as CREATE, ALTER, table/schema DROP, column add/drop/rename, INSERT, UPDATE, DELETE including all-row mutations, TRUNCATE, execute_sql, and apply_migration are authorized for the bound Supabase project and must be executed with tool-result plus read-back proof. Supabase project/account/billing/credential control-plane actions remain denied.',
|
|
558
|
+
requiredSkills: ['mad-db', 'db-safety-guard', 'pipeline-runner', 'context7-docs', REFLECTION_SKILL_NAME, 'honest-mode'],
|
|
559
|
+
appSkillAliases: ['mad-db'],
|
|
560
|
+
lifecycle: ['explicit_invocation', 'single_mission_capability_v2', 'ephemeral_write_profile', 'tool_inventory', 'execute_sql_or_apply_migration', 'read_back_verification', 'close_and_read_only_restore', 'post_route_reflection', 'honest_mode'],
|
|
561
|
+
context7Policy: 'required',
|
|
562
|
+
reasoningPolicy: 'xhigh',
|
|
563
|
+
stopGate: 'mad-db-gate.json',
|
|
564
|
+
cliEntrypoint: 'sks mad-db run|exec|apply-migration|status|close|revoke|doctor',
|
|
565
|
+
examples: ['$MAD-DB public.users legacy_code 컬럼 삭제', '$MAD-DB truncate public.staging_events']
|
|
566
|
+
},
|
|
552
567
|
{
|
|
553
568
|
id: 'MadSKS',
|
|
554
569
|
command: '$MAD-SKS',
|
|
555
570
|
mode: 'MADSKS',
|
|
556
571
|
route: 'explicit scoped permission-widening modifier',
|
|
557
|
-
description: 'Explicit high-risk authorization modifier that can be combined with other $ commands to temporarily open approved target-project scopes such as files, shell, package installs, services, network, Computer Use/browser workflows, generated assets, file permissions, migrations, Supabase MCP DB writes, direct execute SQL, schema cleanup, and normal targeted DB writes for the active invocation, while preserving catastrophic wipe/all-row/project-management, credential-exfiltration, persistent security-weakening, and unrequested fallback safeguards.',
|
|
572
|
+
description: 'Explicit high-risk authorization modifier that can be combined with other $ commands to temporarily open approved target-project scopes such as files, shell, package installs, services, network, Computer Use/browser workflows, generated assets, file permissions, migrations, Supabase MCP DB writes, direct execute SQL, schema cleanup, and normal targeted DB writes for the active invocation, while preserving catastrophic wipe/all-row/project-management, credential-exfiltration, persistent security-weakening, and unrequested fallback safeguards. It is not the first-class MadDB destructive SQL-plane route.',
|
|
558
573
|
requiredSkills: ['mad-sks', 'db-safety-guard', 'pipeline-runner', 'context7-docs', REFLECTION_SKILL_NAME, 'honest-mode'],
|
|
559
|
-
dollarAliases: ['$MAD-DB'],
|
|
560
|
-
appSkillAliases: ['mad-db'],
|
|
561
574
|
lifecycle: ['explicit_invocation', 'auto_sealed_permission_scope', 'scoped_permission_override', 'catastrophic_guard', 'permission_deactivation', 'post_route_reflection', 'honest_mode'],
|
|
562
575
|
context7Policy: 'required',
|
|
563
576
|
reasoningPolicy: 'xhigh',
|
|
564
577
|
stopGate: 'mad-sks-gate.json',
|
|
565
578
|
cliEntrypoint: 'Codex App prompt route only: $MAD-SKS <task>',
|
|
566
|
-
examples: ['$MAD-SKS $Team target project maintenance with package/service/file and DB scopes', '$DB Supabase 점검 $MAD-SKS'
|
|
579
|
+
examples: ['$MAD-SKS $Team target project maintenance with package/service/file and DB scopes', '$DB Supabase 점검 $MAD-SKS']
|
|
567
580
|
},
|
|
568
581
|
{
|
|
569
582
|
id: 'GX',
|
|
@@ -964,6 +977,8 @@ export function routeRequiresSubagents(route, prompt = '') {
|
|
|
964
977
|
return false;
|
|
965
978
|
if (route.id === 'ImageUXReview')
|
|
966
979
|
return false;
|
|
980
|
+
if (route.id === 'MadDB')
|
|
981
|
+
return false;
|
|
967
982
|
if (route.id === 'Research' || route.id === 'AutoResearch')
|
|
968
983
|
return true;
|
|
969
984
|
if (route.id === 'Goal')
|
|
@@ -996,7 +1011,7 @@ export function simpleGitOnlyRouteId(prompt = '') {
|
|
|
996
1011
|
}
|
|
997
1012
|
export function reflectionRequiredForRoute(route) {
|
|
998
1013
|
const id = String(route?.id || route?.mode || route?.route || route || '').replace(/^\$/, '');
|
|
999
|
-
return /^(team|naruto|shadowclone|shadow-clone|kagebunshin|kage-bunshin|qaloop|qa-loop|ppt|imageuxreview|image-ux-review|research|autoresearch|db|database|madsks|mad-sks|gx)$/i.test(id);
|
|
1014
|
+
return /^(team|naruto|shadowclone|shadow-clone|kagebunshin|kage-bunshin|qaloop|qa-loop|ppt|imageuxreview|image-ux-review|research|autoresearch|db|database|madsks|mad-sks|maddb|mad-db|gx)$/i.test(id);
|
|
1000
1015
|
}
|
|
1001
1016
|
export function looksLikeCodeChangingWork(prompt = '') {
|
|
1002
1017
|
const text = String(prompt || '');
|
|
@@ -1038,7 +1053,7 @@ export function routeReasoning(route, prompt = '') {
|
|
|
1038
1053
|
const base = ALLOWED_REASONING_EFFORTS.has(route?.reasoningPolicy) ? route.reasoningPolicy : 'medium';
|
|
1039
1054
|
if (hasFromChatImgSignal(text))
|
|
1040
1055
|
return reasoning('xhigh', 'from_chat_img_image_work_order_analysis');
|
|
1041
|
-
if (/(?:^|\s)sks\s+--mad\b|(?:^|\s)--mad\b|\$MAD-SKS\b|\bmad-sks\b|\bmadsks\b/i.test(text))
|
|
1056
|
+
if (/(?:^|\s)sks\s+--mad\b|(?:^|\s)--mad\b|\$MAD-SKS\b|\$MAD-DB\b|\bmad-sks\b|\bmadsks\b|\bmad-db\b|\bmaddb\b/i.test(text))
|
|
1042
1057
|
return reasoning('xhigh', 'mad_sks_or_mad_launch_default');
|
|
1043
1058
|
if (route?.id === 'Team' || route?.id === 'Naruto')
|
|
1044
1059
|
return teamRouteReasoning(text);
|