ai-engineering-loop 1.14.0 → 1.16.0
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/.agents/{workflows/ai-engineering-loop.md → skills/ai-engineering-loop/SKILL.md} +9 -7
- package/{.gemini → .agents}/skills/ai-engineering-loop/references/loop-deliver.md +1 -0
- package/.agents/skills/ai-engineering-loop/references/loop-hosts.md +48 -0
- package/.agents/skills/ai-engineering-loop/references/loop-maker.md +13 -0
- package/{.gemini → .agents}/skills/ai-engineering-loop/references/loop-review.md +2 -0
- package/.claude/skills/ai-engineering-loop/SKILL.md +10 -11
- package/.claude/skills/ai-engineering-loop/references/loop-deliver.md +1 -0
- package/.claude/skills/ai-engineering-loop/references/loop-hosts.md +48 -0
- package/.claude/skills/ai-engineering-loop/references/loop-maker.md +6 -1
- package/.claude/skills/ai-engineering-loop/references/loop-review.md +2 -0
- package/CHANGELOG.md +83 -0
- package/README.md +54 -6
- package/adapters/README.md +22 -0
- package/adapters/dot/README.md +19 -9
- package/adapters/dot/gitlab.md +49 -0
- package/adapters/dot/mattermost.md +10 -0
- package/adapters/dot/skills/dot-dev-workflow/SKILL.md +27 -5
- package/adapters/dot/sonar.md +90 -0
- package/adapters/generate-adapter/SKILL.md +6 -5
- package/adapters/{dot → quality}/skills/task-impact-inquiry/SKILL.md +10 -7
- package/agents/instruction-only/AGENTS.md +24 -0
- package/bin/ai-engineering-loop.js +320 -34
- package/core/grill-policy.md +1 -1
- package/core/orchestration-model.md +11 -0
- package/core/runtime-safety.md +2 -1
- package/lib/adapter-check.js +102 -0
- package/lib/adapter-detect.js +125 -0
- package/lib/auto-update.js +3 -1
- package/lib/cli-args.js +11 -0
- package/lib/doctor.js +13 -17
- package/lib/gates.js +28 -1
- package/lib/generate-adapter.js +6 -30
- package/lib/host-skill-build.js +25 -15
- package/lib/hosts/agent-formats.js +135 -0
- package/lib/hosts/commands.js +126 -0
- package/lib/hosts/detect.js +56 -0
- package/lib/hosts/doctor.js +77 -0
- package/lib/hosts/redetect.js +60 -0
- package/lib/hosts/registry.js +298 -0
- package/lib/hosts/table.js +43 -0
- package/lib/orchestration.js +39 -6
- package/lib/run-abandon.js +116 -0
- package/lib/studio-server.js +1 -0
- package/lib/sync-hosts.js +156 -112
- package/package.json +2 -3
- package/schemas/report.schema.json +52 -5
- package/.agents/devil-advocate.md +0 -79
- package/.agents/judge.md +0 -58
- package/.claude/commands/ai-engineering-loop.md +0 -11
- package/.gemini/skills/ai-engineering-loop/SKILL.md +0 -61
- package/.gemini/skills/ai-engineering-loop/references/loop-maker.md +0 -8
- package/.gemini/skills/animate/RECIPES.md +0 -324
- package/.gemini/skills/animate/SKILL.md +0 -207
- package/.gemini/skills/animate-expo/RECIPES.md +0 -385
- package/.gemini/skills/animate-expo/SKILL.md +0 -263
- package/.gemini/skills/animation-vocabulary/SKILL.md +0 -181
- package/.gemini/skills/apple-design/SKILL.md +0 -290
- package/.gemini/skills/ask-sonner/API.md +0 -64
- package/.gemini/skills/ask-sonner/SKILL.md +0 -88
- package/.gemini/skills/backend-development/SKILL.md +0 -272
- package/.gemini/skills/backend-development/references/api-design-guidelines.md +0 -95
- package/.gemini/skills/backend-development/references/query-and-database-best-practices.md +0 -53
- package/.gemini/skills/backend-development/references/security-checklist.md +0 -60
- package/.gemini/skills/emil-design-eng/SKILL.md +0 -674
- package/.gemini/skills/find-animation-opportunities/SKILL.md +0 -140
- package/.gemini/skills/improve-animations/AUDIT.md +0 -115
- package/.gemini/skills/improve-animations/PLAN-TEMPLATE.md +0 -73
- package/.gemini/skills/improve-animations/SKILL.md +0 -109
- package/.gemini/skills/mobile-native/SKILL.md +0 -311
- package/.gemini/skills/pick-ui-library/SKILL.md +0 -85
- package/.gemini/skills/prototype/PICKER.md +0 -197
- package/.gemini/skills/prototype/SKILL.md +0 -98
- package/.gemini/skills/review-animations/SKILL.md +0 -120
- package/.gemini/skills/review-animations/STANDARDS.md +0 -187
- package/.gemini/skills/write-swift/SKILL.md +0 -396
- package/.grok/commands/ai-engineering-loop.md +0 -11
- package/.grok/skills/ai-engineering-loop/SKILL.md +0 -62
- package/.grok/skills/ai-engineering-loop/references/loop-commands.md +0 -20
- package/.grok/skills/ai-engineering-loop/references/loop-deliver.md +0 -8
- package/.grok/skills/ai-engineering-loop/references/loop-goal.md +0 -10
- package/.grok/skills/ai-engineering-loop/references/loop-maker.md +0 -8
- package/.grok/skills/ai-engineering-loop/references/loop-review.md +0 -18
- package/.grok/skills/ai-engineering-loop/references/loop-risky.md +0 -9
- package/.grok/skills/ai-engineering-loop/references/loop-verify.md +0 -20
- package/.grok/skills/animate/RECIPES.md +0 -324
- package/.grok/skills/animate/SKILL.md +0 -207
- package/.grok/skills/animate-expo/RECIPES.md +0 -385
- package/.grok/skills/animate-expo/SKILL.md +0 -263
- package/.grok/skills/animation-vocabulary/SKILL.md +0 -181
- package/.grok/skills/apple-design/SKILL.md +0 -290
- package/.grok/skills/ask-sonner/API.md +0 -64
- package/.grok/skills/ask-sonner/SKILL.md +0 -88
- package/.grok/skills/backend-development/SKILL.md +0 -272
- package/.grok/skills/backend-development/references/api-design-guidelines.md +0 -95
- package/.grok/skills/backend-development/references/query-and-database-best-practices.md +0 -53
- package/.grok/skills/backend-development/references/security-checklist.md +0 -60
- package/.grok/skills/emil-design-eng/SKILL.md +0 -674
- package/.grok/skills/find-animation-opportunities/SKILL.md +0 -140
- package/.grok/skills/improve-animations/AUDIT.md +0 -115
- package/.grok/skills/improve-animations/PLAN-TEMPLATE.md +0 -73
- package/.grok/skills/improve-animations/SKILL.md +0 -109
- package/.grok/skills/mobile-native/SKILL.md +0 -311
- package/.grok/skills/pick-ui-library/SKILL.md +0 -85
- package/.grok/skills/prototype/PICKER.md +0 -197
- package/.grok/skills/prototype/SKILL.md +0 -98
- package/.grok/skills/review-animations/SKILL.md +0 -120
- package/.grok/skills/review-animations/STANDARDS.md +0 -187
- package/.grok/skills/write-swift/SKILL.md +0 -396
- /package/{.gemini → .agents}/skills/ai-engineering-loop/references/loop-commands.md +0 -0
- /package/{.gemini → .agents}/skills/ai-engineering-loop/references/loop-goal.md +0 -0
- /package/{.gemini → .agents}/skills/ai-engineering-loop/references/loop-risky.md +0 -0
- /package/{.gemini → .agents}/skills/ai-engineering-loop/references/loop-verify.md +0 -0
|
@@ -18,12 +18,16 @@ const {
|
|
|
18
18
|
homeDir,
|
|
19
19
|
applyHostSync,
|
|
20
20
|
hostSkillStatus,
|
|
21
|
+
resolveHostSelection,
|
|
22
|
+
planHostSync,
|
|
21
23
|
formatHostSyncReport
|
|
22
24
|
} = require('../lib/sync-hosts.js');
|
|
25
|
+
const hostCommands = require('../lib/hosts/commands.js');
|
|
23
26
|
const { listDocs, readDoc } = require('../lib/docs.js');
|
|
24
27
|
const {
|
|
25
28
|
wantsCommandHelp,
|
|
26
29
|
unknownWriterArgs,
|
|
30
|
+
effectiveCommand,
|
|
27
31
|
commandHelp
|
|
28
32
|
} = require('../lib/cli-args.js');
|
|
29
33
|
const {
|
|
@@ -32,6 +36,8 @@ const {
|
|
|
32
36
|
formatGenerateAdapterReport,
|
|
33
37
|
parseTypeArg
|
|
34
38
|
} = require('../lib/generate-adapter.js');
|
|
39
|
+
const { detectAdapter } = require('../lib/adapter-detect.js');
|
|
40
|
+
const { checkAdapter } = require('../lib/adapter-check.js');
|
|
35
41
|
const {
|
|
36
42
|
writeWorkflowOverlay,
|
|
37
43
|
formatGenerateWorkflowReport,
|
|
@@ -545,26 +551,12 @@ function analyzeRepository(rootDir) {
|
|
|
545
551
|
discovery.evidence.push('Manifest: pubspec.yaml');
|
|
546
552
|
}
|
|
547
553
|
|
|
548
|
-
// 3. Adapter / CI Detection
|
|
549
|
-
|
|
550
|
-
|
|
551
|
-
|
|
552
|
-
|
|
553
|
-
|
|
554
|
-
discovery.adapter.type = 'gitlab';
|
|
555
|
-
}
|
|
556
|
-
|
|
557
|
-
// Check git remote if git folder exists
|
|
558
|
-
const gitConfigStr = readFileSafe(path.join(rootDir, '.git', 'config'));
|
|
559
|
-
if (gitConfigStr) {
|
|
560
|
-
const urlMatch = gitConfigStr.match(/url\s*=\s*(.*)/);
|
|
561
|
-
if (urlMatch) {
|
|
562
|
-
discovery.adapter.repoSlug = urlMatch[1].trim();
|
|
563
|
-
if (discovery.adapter.repoSlug.includes('gitlab.dot.co.id')) {
|
|
564
|
-
discovery.adapter.type = 'dot';
|
|
565
|
-
}
|
|
566
|
-
}
|
|
567
|
-
}
|
|
554
|
+
// 3. Adapter / CI Detection (shared with generate-adapter and Studio, #62)
|
|
555
|
+
const detected = detectAdapter(rootDir);
|
|
556
|
+
discovery.adapter.type = detected.type;
|
|
557
|
+
discovery.adapter.ciProvider = detected.ciProvider;
|
|
558
|
+
discovery.adapter.repoSlug = detected.remote || null;
|
|
559
|
+
discovery.evidence.push(`Adapter: ${detected.type} (${detected.reason})`);
|
|
568
560
|
|
|
569
561
|
return discovery;
|
|
570
562
|
}
|
|
@@ -909,6 +901,12 @@ function statusJson() {
|
|
|
909
901
|
result.contextVersion = metadata.contextVersion || null;
|
|
910
902
|
result.projectProfile = metadata.projectProfile || null;
|
|
911
903
|
result.freshness = { status: drift.status, reason: drift.reason, level: drift.level || null };
|
|
904
|
+
const activeRun = currentRunSafe(CWD);
|
|
905
|
+
if (activeRun && !TERMINAL_STATES.has(activeRun.state)) {
|
|
906
|
+
const info = require('../lib/run-abandon.js').staleRunInfo(activeRun);
|
|
907
|
+
result.activeRun = { runId: activeRun.runId, state: activeRun.state, updatedAt: activeRun.updatedAt || null, stale: info.stale, idleDays: info.idleDays };
|
|
908
|
+
if (info.stale) result.activeRun.hint = require('../lib/run-abandon.js').abandonCommand(activeRun.runId);
|
|
909
|
+
}
|
|
912
910
|
}
|
|
913
911
|
}
|
|
914
912
|
return result;
|
|
@@ -950,6 +948,13 @@ function handleStatus() {
|
|
|
950
948
|
console.log(`- Living Freshness: \x1b[32m${drift.status}\x1b[0m (${drift.reason})`);
|
|
951
949
|
console.log(`- Context Files: ${REQUIRED_FILES.length}/${REQUIRED_FILES.length} verified (including metadata.json, glossary.md, adrs/)`);
|
|
952
950
|
|
|
951
|
+
const activeRun = currentRunSafe(CWD);
|
|
952
|
+
if (activeRun && !TERMINAL_STATES.has(activeRun.state)) {
|
|
953
|
+
console.log(`- Active run: ${activeRun.runId} (${activeRun.state})`);
|
|
954
|
+
const hint = staleRunHint(activeRun);
|
|
955
|
+
if (hint) log.warn(` ${hint}`);
|
|
956
|
+
}
|
|
957
|
+
|
|
953
958
|
const hosts = hostSkillStatus({ home: homeDir(), version: VERSION });
|
|
954
959
|
if (hosts.stale) {
|
|
955
960
|
log.warn(`- Host skills: STALE (${hosts.staleFiles} file(s) behind package v${VERSION}${hosts.installedVersion ? `, installed v${hosts.installedVersion}` : ''})`);
|
|
@@ -1030,6 +1035,168 @@ function handleSyncHosts() {
|
|
|
1030
1035
|
console.log(formatHostSyncReport(results, { version: VERSION, dryRun }));
|
|
1031
1036
|
}
|
|
1032
1037
|
|
|
1038
|
+
function hostSyncOrExit(selection, { dryRun = false } = {}) {
|
|
1039
|
+
try {
|
|
1040
|
+
return applyHostSync({ home: homeDir(), hosts: selection, dryRun, version: VERSION });
|
|
1041
|
+
} catch (err) {
|
|
1042
|
+
log.error(`${err.code || 'SYNC_FAILED'}: ${err.message}`);
|
|
1043
|
+
process.exit(2);
|
|
1044
|
+
}
|
|
1045
|
+
}
|
|
1046
|
+
|
|
1047
|
+
function readTtyLine(question, fallback) {
|
|
1048
|
+
process.stdout.write(question);
|
|
1049
|
+
try {
|
|
1050
|
+
const buffer = Buffer.alloc(64);
|
|
1051
|
+
const bytes = fs.readSync(0, buffer, 0, buffer.length, null);
|
|
1052
|
+
return buffer.toString('utf8', 0, bytes).trim().toLowerCase() || fallback;
|
|
1053
|
+
} catch {
|
|
1054
|
+
return fallback;
|
|
1055
|
+
}
|
|
1056
|
+
}
|
|
1057
|
+
|
|
1058
|
+
// `hosts list|doctor|enable|disable` (issues #49, #52).
|
|
1059
|
+
function handleHosts() {
|
|
1060
|
+
const rest = process.argv.slice(3).filter((arg) => !arg.startsWith('-'));
|
|
1061
|
+
const sub = rest[0] || 'list';
|
|
1062
|
+
const json = process.argv.includes('--json');
|
|
1063
|
+
const home = homeDir();
|
|
1064
|
+
const selection = resolveHostSelection(home);
|
|
1065
|
+
if (sub === 'list') {
|
|
1066
|
+
const rows = hostCommands.listHosts({ home, selection });
|
|
1067
|
+
if (json) {
|
|
1068
|
+
console.log(JSON.stringify({ ok: true, source: selection.source, mode: selection.mode, hosts: rows }, null, 2));
|
|
1069
|
+
return;
|
|
1070
|
+
}
|
|
1071
|
+
console.log(hostCommands.formatHostsList(rows, selection));
|
|
1072
|
+
if (selection.source !== 'config') console.log('Run "ai-engineering-loop setup" to choose hosts explicitly.');
|
|
1073
|
+
return;
|
|
1074
|
+
}
|
|
1075
|
+
if (sub === 'doctor') {
|
|
1076
|
+
const { hostsDoctor, formatHostsDoctor } = require('../lib/hosts/doctor.js');
|
|
1077
|
+
const report = hostsDoctor({ home, selection, plan: planHostSync({ home, hosts: selection }) });
|
|
1078
|
+
console.log(json ? JSON.stringify(report, null, 2) : formatHostsDoctor(report));
|
|
1079
|
+
if (!report.ok) process.exitCode = 1;
|
|
1080
|
+
return;
|
|
1081
|
+
}
|
|
1082
|
+
if (sub !== 'enable' && sub !== 'disable') {
|
|
1083
|
+
log.error(`Unknown hosts subcommand "${sub}". Use: list, doctor, enable <id...>, disable <id...>. Nothing was written.`);
|
|
1084
|
+
process.exit(2);
|
|
1085
|
+
}
|
|
1086
|
+
let ids;
|
|
1087
|
+
try {
|
|
1088
|
+
ids = hostCommands.parseHostIds(rest.slice(1));
|
|
1089
|
+
} catch (err) {
|
|
1090
|
+
log.error(`${err.code}: ${err.message}. Nothing was written.`);
|
|
1091
|
+
process.exit(2);
|
|
1092
|
+
}
|
|
1093
|
+
if (!ids.length) {
|
|
1094
|
+
log.error(`hosts ${sub} needs at least one host id. Nothing was written.`);
|
|
1095
|
+
process.exit(2);
|
|
1096
|
+
}
|
|
1097
|
+
const next = hostCommands.changeHosts({ home, selection, [sub]: ids });
|
|
1098
|
+
const results = hostSyncOrExit(next);
|
|
1099
|
+
if (json) {
|
|
1100
|
+
console.log(JSON.stringify({ ok: true, [sub === 'enable' ? 'enabled' : 'disabled']: ids, hosts: [...next.enabled], files: results.map(({ content, ...item }) => item) }, null, 2));
|
|
1101
|
+
return;
|
|
1102
|
+
}
|
|
1103
|
+
log.info(`AI Engineering Loop — hosts ${sub} ${ids.join(', ')}`);
|
|
1104
|
+
console.log(formatHostSyncReport(results, { version: VERSION }));
|
|
1105
|
+
}
|
|
1106
|
+
|
|
1107
|
+
// `setup` (issue #49): the one-time, explicit replacement for a postinstall
|
|
1108
|
+
// hook. Detects agents (config folder or PATH binary, never executed), asks,
|
|
1109
|
+
// then syncs. Without a TTY it only prints the plan unless --yes is given.
|
|
1110
|
+
function handleSetup() {
|
|
1111
|
+
const home = homeDir();
|
|
1112
|
+
const json = process.argv.includes('--json');
|
|
1113
|
+
const yes = process.argv.includes('--yes');
|
|
1114
|
+
const hostsArg = argValue('--hosts');
|
|
1115
|
+
const modeArg = argValue('--mode');
|
|
1116
|
+
if (modeArg && !['auto', 'manual'].includes(modeArg)) {
|
|
1117
|
+
log.error('--mode must be auto or manual. Nothing was written.');
|
|
1118
|
+
process.exit(2);
|
|
1119
|
+
}
|
|
1120
|
+
const selection = resolveHostSelection(home);
|
|
1121
|
+
const rows = hostCommands.listHosts({ home, selection });
|
|
1122
|
+
let ids;
|
|
1123
|
+
try {
|
|
1124
|
+
ids = hostsArg ? hostCommands.parseHostIds(hostsArg) : rows.filter((row) => row.detected).map((row) => row.id);
|
|
1125
|
+
} catch (err) {
|
|
1126
|
+
log.error(`${err.code}: ${err.message}. Nothing was written.`);
|
|
1127
|
+
process.exit(2);
|
|
1128
|
+
}
|
|
1129
|
+
const interactive = process.stdin.isTTY && process.stdout.isTTY && !json && autoUpdate.autoUpdateAllowedHere();
|
|
1130
|
+
if (!yes && !interactive) {
|
|
1131
|
+
const plan = { ok: true, wrote: false, detected: rows.filter((row) => row.detected).map((row) => row.id), wouldEnable: ids };
|
|
1132
|
+
if (json) console.log(JSON.stringify(plan, null, 2));
|
|
1133
|
+
else {
|
|
1134
|
+
console.log(hostCommands.formatHostsList(rows, selection));
|
|
1135
|
+
console.log(`Would enable: ${ids.join(', ') || '(none detected)'}`);
|
|
1136
|
+
console.log('Nothing was written. Run "ai-engineering-loop setup --yes" (optionally --hosts a,b --mode auto|manual) to install.');
|
|
1137
|
+
}
|
|
1138
|
+
return;
|
|
1139
|
+
}
|
|
1140
|
+
let mode = hostCommands.setupMode({ modeArg, interactive: false });
|
|
1141
|
+
if (!yes) {
|
|
1142
|
+
console.log(hostCommands.formatHostsList(rows, selection));
|
|
1143
|
+
if (!ids.length) {
|
|
1144
|
+
console.log('No AI agent detected. Install one, or pick hosts with "setup --hosts <id,...>".');
|
|
1145
|
+
return;
|
|
1146
|
+
}
|
|
1147
|
+
const ok = readTtyLine(`Install the loop for ${ids.join(', ')}? [Y/n] `, 'y');
|
|
1148
|
+
if (!['y', 'yes'].includes(ok)) {
|
|
1149
|
+
console.log('Nothing was written.');
|
|
1150
|
+
return;
|
|
1151
|
+
}
|
|
1152
|
+
if (!modeArg) {
|
|
1153
|
+
mode = hostCommands.setupMode({ interactive: true, answer: readTtyLine(hostCommands.SETUP_MODE_QUESTION, '') });
|
|
1154
|
+
}
|
|
1155
|
+
}
|
|
1156
|
+
const next = hostCommands.applySetup({ home, ids, mode });
|
|
1157
|
+
const results = hostSyncOrExit(next);
|
|
1158
|
+
if (json) {
|
|
1159
|
+
console.log(JSON.stringify({ ok: true, wrote: true, enabled: [...next.enabled], mode: next.mode, files: results.map(({ content, ...item }) => item) }, null, 2));
|
|
1160
|
+
return;
|
|
1161
|
+
}
|
|
1162
|
+
log.info(`AI Engineering Loop — setup (${next.enabled.size} host${next.enabled.size === 1 ? '' : 's'}, mode ${next.mode})`);
|
|
1163
|
+
console.log(formatHostSyncReport(results, { version: VERSION }));
|
|
1164
|
+
}
|
|
1165
|
+
|
|
1166
|
+
// First-run hint (issue #49): no postinstall writes to HOME, so the first
|
|
1167
|
+
// interactive CLI use suggests `setup` once a day until hosts are chosen.
|
|
1168
|
+
function maybeSetupHint() {
|
|
1169
|
+
try {
|
|
1170
|
+
if (process.env.AEL_NO_SETUP_HINT === '1' || process.argv.includes('--json')) return;
|
|
1171
|
+
if (!process.stderr.isTTY || !autoUpdate.autoUpdateAllowedHere()) return;
|
|
1172
|
+
const home = homeDir();
|
|
1173
|
+
if (hostCommands.savedHosts(home)) return;
|
|
1174
|
+
const userState = require('../lib/user-state.js');
|
|
1175
|
+
const stampFile = userState.userStatePath('setup-hint.json', home);
|
|
1176
|
+
const last = userState.readJsonFile(stampFile, {});
|
|
1177
|
+
if (last.at && Date.now() - Date.parse(last.at) < 24 * 60 * 60 * 1000) return;
|
|
1178
|
+
const rows = hostCommands.listHosts({ home, selection: resolveHostSelection(home) });
|
|
1179
|
+
const detected = rows.filter((row) => row.detected).map((row) => row.id);
|
|
1180
|
+
if (!detected.length) return;
|
|
1181
|
+
userState.writeJsonFile(stampFile, { at: new Date().toISOString() });
|
|
1182
|
+
process.stderr.write(`ai-engineering-loop: detected ${detected.join(', ')}. Run "ai-engineering-loop setup" once to install the loop for them (silence: AEL_NO_SETUP_HINT=1).\n`);
|
|
1183
|
+
} catch {
|
|
1184
|
+
// A hint never breaks a command.
|
|
1185
|
+
}
|
|
1186
|
+
}
|
|
1187
|
+
|
|
1188
|
+
// New agents after setup (issue #52): `status` prints a short, rate-limited
|
|
1189
|
+
// notice in manual mode; auto mode is handled by the background `update auto`.
|
|
1190
|
+
function maybeNewHostsNotice() {
|
|
1191
|
+
try {
|
|
1192
|
+
if (process.argv.includes('--json') || !process.stderr.isTTY || !autoUpdate.autoUpdateAllowedHere()) return;
|
|
1193
|
+
const message = require('../lib/hosts/redetect.js').newHostsNotice({ home: homeDir() });
|
|
1194
|
+
if (message) process.stderr.write(`${message}\n`);
|
|
1195
|
+
} catch {
|
|
1196
|
+
// A notice never breaks a command.
|
|
1197
|
+
}
|
|
1198
|
+
}
|
|
1199
|
+
|
|
1033
1200
|
function handleGenerateAdapter() {
|
|
1034
1201
|
log.info('AI Engineering Loop — Generate delivery adapter (generate-adapter)');
|
|
1035
1202
|
const hints = detectAdapterHints(CWD);
|
|
@@ -1085,7 +1252,72 @@ function workingTreePaths(rootDir) {
|
|
|
1085
1252
|
}
|
|
1086
1253
|
|
|
1087
1254
|
// Command: run
|
|
1255
|
+
// `run --abandon [--run <id>] --reason <text> --by <name> [--yes] [--json]`
|
|
1256
|
+
// (issue #46). ASSISTED asks for confirmation (or --yes); UNATTENDED runs are
|
|
1257
|
+
// refused unless a human passes --yes. Terminal runs exit 2.
|
|
1258
|
+
function handleRunAbandon() {
|
|
1259
|
+
const json = process.argv.includes('--json');
|
|
1260
|
+
const yes = process.argv.includes('--yes');
|
|
1261
|
+
const runAbandon = require('../lib/run-abandon.js');
|
|
1262
|
+
const fail = (err) => {
|
|
1263
|
+
if (json) console.log(JSON.stringify({ ok: false, code: err.code || 'ABANDON_FAILED', error: err.message }));
|
|
1264
|
+
else log.error(err.message);
|
|
1265
|
+
process.exit(err.exitCode || 1);
|
|
1266
|
+
};
|
|
1267
|
+
try {
|
|
1268
|
+
const plan = runAbandon.planAbandon(CWD, { runId: argValue('--run'), reason: argValue('--reason'), by: argValue('--by') });
|
|
1269
|
+
const { run } = plan;
|
|
1270
|
+
if (run.mode === 'UNATTENDED' && !yes) {
|
|
1271
|
+
throw Object.assign(new Error(`Run ${run.runId} is UNATTENDED. Agents must not abandon it; a human can re-run with --yes.`), { code: 'UNATTENDED_REQUIRES_HUMAN', exitCode: 2 });
|
|
1272
|
+
}
|
|
1273
|
+
if (!yes) {
|
|
1274
|
+
if (json || !process.stdin.isTTY || !process.stdout.isTTY) {
|
|
1275
|
+
throw Object.assign(new Error(`Confirm abandoning run ${run.runId} (${run.state}) by re-running with --yes.`), { code: 'CONFIRMATION_REQUIRED' });
|
|
1276
|
+
}
|
|
1277
|
+
process.stdout.write(`Abandon run ${run.runId} (${run.state}, ${run.task ? `"${String(run.task).slice(0, 60)}"` : 'no task'})? Run data is kept. [y/N] `);
|
|
1278
|
+
const buffer = Buffer.alloc(16);
|
|
1279
|
+
let answer = '';
|
|
1280
|
+
try {
|
|
1281
|
+
answer = buffer.toString('utf8', 0, fs.readSync(0, buffer, 0, buffer.length, null)).trim().toLowerCase();
|
|
1282
|
+
} catch {
|
|
1283
|
+
answer = '';
|
|
1284
|
+
}
|
|
1285
|
+
if (!['y', 'yes'].includes(answer)) {
|
|
1286
|
+
console.log('Not abandoned.');
|
|
1287
|
+
process.exit(1);
|
|
1288
|
+
}
|
|
1289
|
+
}
|
|
1290
|
+
const result = runAbandon.abandonRun(CWD, { runId: run.runId, reason: plan.reason, by: plan.actor });
|
|
1291
|
+
if (json) console.log(JSON.stringify({ ok: true, ...result }));
|
|
1292
|
+
else {
|
|
1293
|
+
log.success(`✓ Run ${result.runId} abandoned (${result.from} -> CANCELLED) by ${result.by}: ${result.reason}`);
|
|
1294
|
+
console.log(result.released
|
|
1295
|
+
? '- current.json released; refresh and auto-update will process this project again.'
|
|
1296
|
+
: `- current run unchanged${result.current ? ` (${result.current})` : ''}.`);
|
|
1297
|
+
}
|
|
1298
|
+
} catch (err) {
|
|
1299
|
+
fail(err);
|
|
1300
|
+
}
|
|
1301
|
+
}
|
|
1302
|
+
|
|
1303
|
+
// One-line hint for an active run that has not moved for STALE_RUN_DAYS.
|
|
1304
|
+
function staleRunHint(run) {
|
|
1305
|
+
const runAbandon = require('../lib/run-abandon.js');
|
|
1306
|
+
const info = runAbandon.staleRunInfo(run);
|
|
1307
|
+
if (!info.stale) return null;
|
|
1308
|
+
return `Run ${run.runId} has been ${run.state} for ${info.idleDays} day(s). If it is abandoned: ${runAbandon.abandonCommand(run.runId)}`;
|
|
1309
|
+
}
|
|
1310
|
+
|
|
1311
|
+
function currentRunSafe(dir) {
|
|
1312
|
+
try {
|
|
1313
|
+
return getCurrentRun(dir);
|
|
1314
|
+
} catch {
|
|
1315
|
+
return null;
|
|
1316
|
+
}
|
|
1317
|
+
}
|
|
1318
|
+
|
|
1088
1319
|
function handleRun() {
|
|
1320
|
+
if (process.argv.includes('--abandon')) return handleRunAbandon();
|
|
1089
1321
|
const runJson = process.argv.includes('--json');
|
|
1090
1322
|
const realLog = console.log;
|
|
1091
1323
|
if (runJson) console.log = () => {};
|
|
@@ -1228,7 +1460,8 @@ function handleRun() {
|
|
|
1228
1460
|
console.log('1. Grill if needed (docs grill-policy), then freeze a failure-table Goal Contract');
|
|
1229
1461
|
if (runResult.run.mode === RUN_MODES.REPORT_ONLY) {
|
|
1230
1462
|
console.log('2. Analyze evidence without changing repository files');
|
|
1231
|
-
console.log('3. Write report.json
|
|
1463
|
+
console.log('3. Write report.json; if the host has a subagent tool, have ONE fresh devil-advocate challenge it (no Judge), and record `review` in report.json (loop-maker)');
|
|
1464
|
+
console.log('4. Finish with `ai-engineering-loop gate report`');
|
|
1232
1465
|
} else {
|
|
1233
1466
|
console.log('2. Root Cause Analysis (docs root-cause-analysis) & Plan');
|
|
1234
1467
|
console.log('3. Maker TDD at named seams (docs tdd-policy)');
|
|
@@ -1264,7 +1497,7 @@ function handleRun() {
|
|
|
1264
1497
|
} else {
|
|
1265
1498
|
console.log(`- Fallback: CONTEXT_ISOLATION_ONLY (${grok.reason})`);
|
|
1266
1499
|
}
|
|
1267
|
-
console.log('- Skill: .
|
|
1500
|
+
console.log('- Skill: .agents/skills/ai-engineering-loop/SKILL.md (spawn rules: references/loop-hosts.md, Grok CLI section)');
|
|
1268
1501
|
}
|
|
1269
1502
|
|
|
1270
1503
|
const claudeSkill = path.join(CWD, '.claude', 'skills', 'ai-engineering-loop', 'SKILL.md');
|
|
@@ -1279,14 +1512,14 @@ function handleRun() {
|
|
|
1279
1512
|
console.log('- Skill: .claude/skills/ai-engineering-loop/SKILL.md');
|
|
1280
1513
|
}
|
|
1281
1514
|
|
|
1282
|
-
const
|
|
1283
|
-
if (fs.existsSync(
|
|
1515
|
+
const sharedSkill = path.join(CWD, '.agents', 'skills', 'ai-engineering-loop', 'SKILL.md');
|
|
1516
|
+
if (fs.existsSync(sharedSkill)) {
|
|
1284
1517
|
console.log('------------------------------------------------------------');
|
|
1285
|
-
log.bold('Antigravity
|
|
1286
|
-
console.log('-
|
|
1518
|
+
log.bold('Shared .agents/skills hosts (Codex, Gemini CLI, Antigravity, Cursor, Copilot, ...):');
|
|
1519
|
+
console.log('- Same skill on every host; spawn rules per host: references/loop-hosts.md');
|
|
1287
1520
|
console.log('- Devil\'s Advocate: 8 tool calls, diff file, skip css');
|
|
1288
1521
|
console.log('- Judge: 4 tool calls, ledger + contract only');
|
|
1289
|
-
console.log('-
|
|
1522
|
+
console.log('- Skill: .agents/skills/ai-engineering-loop/SKILL.md');
|
|
1290
1523
|
}
|
|
1291
1524
|
console.log('------------------------------------------------------------\n');
|
|
1292
1525
|
}
|
|
@@ -2001,6 +2234,8 @@ function handleState() {
|
|
|
2001
2234
|
console.log(`State: ${run.state}`);
|
|
2002
2235
|
console.log(`Iteration: ${run.iteration}`);
|
|
2003
2236
|
console.log(`Artifacts: ${Object.keys(run.artifacts).join(', ') || 'none'}`);
|
|
2237
|
+
const hint = staleRunHint(run);
|
|
2238
|
+
if (hint) log.warn(hint);
|
|
2004
2239
|
}
|
|
2005
2240
|
} catch (err) {
|
|
2006
2241
|
if (json) console.log(JSON.stringify({ ok: false, code: err.code || 'STATE_FAILED', error: err.message }));
|
|
@@ -2011,12 +2246,12 @@ function handleState() {
|
|
|
2011
2246
|
|
|
2012
2247
|
function handleDoctor() {
|
|
2013
2248
|
const json = process.argv.includes('--json');
|
|
2014
|
-
const result = runDoctor(path.join(__dirname, '..'));
|
|
2249
|
+
const result = runDoctor(path.join(__dirname, '..'), { projectDir: CWD });
|
|
2015
2250
|
if (json) console.log(JSON.stringify(result));
|
|
2016
2251
|
else {
|
|
2017
2252
|
console.log(`AI Engineering Loop doctor: ${result.ok ? 'READY' : 'NOT READY'}`);
|
|
2018
2253
|
for (const check of result.checks) {
|
|
2019
|
-
console.log(`${check.passed ? '✓' : '✗'} ${check.id}: ${check.detail}`);
|
|
2254
|
+
console.log(`${check.passed ? '✓' : check.level === 'warn' ? '!' : '✗'} ${check.id}: ${check.detail}`);
|
|
2020
2255
|
}
|
|
2021
2256
|
}
|
|
2022
2257
|
if (!result.ok) process.exit(1);
|
|
@@ -2044,6 +2279,7 @@ function handleSafeRefresh() {
|
|
|
2044
2279
|
const current = getCurrentRun(CWD);
|
|
2045
2280
|
if (current && !TERMINAL_STATES.has(current.state)) {
|
|
2046
2281
|
log.warn(`Run ${current.runId} is active (${current.state}); skipping refresh until it finishes.`);
|
|
2282
|
+
log.dim(`If that run is abandoned, close it first: ${require('../lib/run-abandon.js').abandonCommand(current.runId)}`);
|
|
2047
2283
|
process.exit(3);
|
|
2048
2284
|
}
|
|
2049
2285
|
const legacy = !fs.existsSync(path.join(CONTEXT_DIR, GENERATED_HASHES_FILE));
|
|
@@ -2190,6 +2426,17 @@ function handleUpdateAuto() {
|
|
|
2190
2426
|
// Best effort.
|
|
2191
2427
|
}
|
|
2192
2428
|
const decision = autoUpdate.decideUpdate({ current: VERSION, distTags: cache.distTags, ...policy });
|
|
2429
|
+
// Agents installed after setup (#52): auto mode enables and syncs them here.
|
|
2430
|
+
let hostsResult = { added: [] };
|
|
2431
|
+
try {
|
|
2432
|
+
hostsResult = require('../lib/hosts/redetect.js').autoEnableNewHosts({ home: homeDir(), scope });
|
|
2433
|
+
if (hostsResult.added.length && decision.action !== 'apply') {
|
|
2434
|
+
const synced = applyHostSync({ home: homeDir(), hosts: resolveHostSelection(homeDir()), version: VERSION });
|
|
2435
|
+
hostsResult = { ...hostsResult, synced: synced.filter((item) => item.action === 'copy').length };
|
|
2436
|
+
}
|
|
2437
|
+
} catch (err) {
|
|
2438
|
+
hostsResult = { added: [], error: err.message };
|
|
2439
|
+
}
|
|
2193
2440
|
const projects = projectRegistry.listProjects().projects;
|
|
2194
2441
|
let result = null;
|
|
2195
2442
|
if (decision.action === 'apply') {
|
|
@@ -2205,7 +2452,10 @@ function handleUpdateAuto() {
|
|
|
2205
2452
|
} else {
|
|
2206
2453
|
result = autoUpdate.processPending({ projects, readCurrentRun: readCurrentRunFor });
|
|
2207
2454
|
}
|
|
2208
|
-
|
|
2455
|
+
const held = result && Array.isArray(result.pending) && result.pending.length
|
|
2456
|
+
? { pendingHint: `Projects skipped because a run is active. If a run is abandoned, run in that project: ${require('../lib/run-abandon.js').abandonCommand()}` }
|
|
2457
|
+
: {};
|
|
2458
|
+
out({ ok: true, policy: policy.policy, policySource: policy.source, scope, offline: cache.error || null, decision, result, hosts: hostsResult, ...held });
|
|
2209
2459
|
cacheUpdateHighlights({ ...cache, scope }, decision, packageName).catch(() => {});
|
|
2210
2460
|
}
|
|
2211
2461
|
|
|
@@ -2258,7 +2508,12 @@ function handleProjects() {
|
|
|
2258
2508
|
if (json) console.log(JSON.stringify({ ok: true, action, ...result }));
|
|
2259
2509
|
else if (action === 'list') {
|
|
2260
2510
|
if (!result.projects.length) console.log('No registered projects. They are added by init/run/status/refresh, or `projects add|scan`.');
|
|
2261
|
-
for (const p of result.projects)
|
|
2511
|
+
for (const p of result.projects) {
|
|
2512
|
+
const held = currentRunSafe(p.path);
|
|
2513
|
+
const info = held ? require('../lib/run-abandon.js').staleRunInfo(held) : { stale: false };
|
|
2514
|
+
const tag = info.stale ? `; held by stale run ${held.runId} (${held.state}, ${info.idleDays}d): run --abandon there` : '';
|
|
2515
|
+
console.log(`- ${p.path} (context ${p.contextVersion || '?'}, seen ${p.lastSeen}${tag})`);
|
|
2516
|
+
}
|
|
2262
2517
|
} else console.log(JSON.stringify(result, null, 2));
|
|
2263
2518
|
} catch (err) {
|
|
2264
2519
|
if (json) console.log(JSON.stringify({ ok: false, code: err.code || 'PROJECTS_FAILED', error: err.message }));
|
|
@@ -2961,6 +3216,18 @@ function handleDeliver() {
|
|
|
2961
3216
|
if (approve && !by) {
|
|
2962
3217
|
throw Object.assign(new Error('--approve requires --by <name> (the approving human).'), { code: 'INVALID_USAGE' });
|
|
2963
3218
|
}
|
|
3219
|
+
// Adapter preflight (#63): UNATTENDED stops before anything is written;
|
|
3220
|
+
// ASSISTED only warns, because a human is approving this delivery.
|
|
3221
|
+
// A configured-vs-detected mismatch can be deliberate, so it never blocks.
|
|
3222
|
+
const adapterProblems = checkAdapter(CWD).filter((item) => !item.passed);
|
|
3223
|
+
const blocking = adapterProblems.filter((item) => item.id !== 'adapter-detected');
|
|
3224
|
+
if (mode === RUN_MODES.UNATTENDED && blocking.length) {
|
|
3225
|
+
throw Object.assign(
|
|
3226
|
+
new Error(`Adapter preflight failed; nothing was written. ${blocking.map((item) => `${item.id}: ${item.detail}`).join('; ')}`),
|
|
3227
|
+
{ code: 'ADAPTER_PREFLIGHT_FAILED' }
|
|
3228
|
+
);
|
|
3229
|
+
}
|
|
3230
|
+
for (const item of adapterProblems) console.error(`! ${item.id}: ${item.detail}`);
|
|
2964
3231
|
const approved = [];
|
|
2965
3232
|
if (approve && run.workflow) {
|
|
2966
3233
|
const workflow = workflowStatus(CWD, { runId });
|
|
@@ -3046,6 +3313,9 @@ Commands:
|
|
|
3046
3313
|
--recipe <id> bind an immutable compiled workflow plan
|
|
3047
3314
|
--tier T0|T1|T2 --tier-reason <text> override the automatic tier (recorded)
|
|
3048
3315
|
--json print runId, tier, tier reasons, and subagent count only
|
|
3316
|
+
--abandon [--run <id>] --reason <text> --by <name> [--yes] [--json]
|
|
3317
|
+
close a stale run as CANCELLED (kept on disk) and release current.json;
|
|
3318
|
+
ASSISTED asks to confirm, UNATTENDED needs --yes from a human
|
|
3049
3319
|
state Inspect the current run ledger
|
|
3050
3320
|
--run <id> target a non-current run
|
|
3051
3321
|
--json print the complete machine-readable ledger
|
|
@@ -3162,7 +3432,13 @@ Commands:
|
|
|
3162
3432
|
create [--run <id>] [--audience developer|agent|auditor]
|
|
3163
3433
|
[--output <repository-relative-file>|-] [--json]
|
|
3164
3434
|
inspect | brief <repository-relative-file> [--json]
|
|
3165
|
-
|
|
3435
|
+
setup One-time: detect AI agents (config folder or PATH binary, never run) and install for them
|
|
3436
|
+
--yes skip the prompt --hosts <id,...> pick hosts --json
|
|
3437
|
+
--mode auto|manual (prompt suggests auto; --yes / no TTY default to manual)
|
|
3438
|
+
Without a TTY and without --yes it only prints the plan
|
|
3439
|
+
hosts Show or change which AI agent hosts get the loop
|
|
3440
|
+
list [--json] | doctor [--json] | enable <id...> | disable <id...> (removes unedited AEL files)
|
|
3441
|
+
sync-hosts Install the loop skill and reviewer agents into every detected AI agent host
|
|
3166
3442
|
(only hosts that already exist; DOT skills only if already installed)
|
|
3167
3443
|
--dry-run print the plan without writing
|
|
3168
3444
|
--pack <list> core (always), backend, design, quality; saved for later syncs
|
|
@@ -3194,14 +3470,15 @@ const command = args[0] || 'init';
|
|
|
3194
3470
|
// Side-effect guard (issue #25): help never executes a command, and writer
|
|
3195
3471
|
// commands reject unknown flags before touching any file.
|
|
3196
3472
|
if (!['-h', '--help', '-v', '--version'].includes(command)) {
|
|
3197
|
-
|
|
3473
|
+
const checked = effectiveCommand(command, args.slice(1));
|
|
3474
|
+
if (wantsCommandHelp(checked, args.slice(1))) {
|
|
3198
3475
|
const block = commandHelp(helpText(), command);
|
|
3199
3476
|
console.log(block
|
|
3200
3477
|
? `Usage: npx ai-engineering-loop ${block.trim()}`
|
|
3201
3478
|
: helpText());
|
|
3202
3479
|
process.exit(0);
|
|
3203
3480
|
}
|
|
3204
|
-
const unknown = unknownWriterArgs(
|
|
3481
|
+
const unknown = unknownWriterArgs(checked, args.slice(1));
|
|
3205
3482
|
if (unknown.length > 0) {
|
|
3206
3483
|
log.error(`Unknown argument(s) for "${command}": ${unknown.join(' ')}. Nothing was written.`);
|
|
3207
3484
|
console.log(`Run "npx ai-engineering-loop ${command} --help" for usage.`);
|
|
@@ -3217,6 +3494,9 @@ if (['init', 'run', 'status', 'refresh'].includes(command)) {
|
|
|
3217
3494
|
maybeAutoUpdate();
|
|
3218
3495
|
}
|
|
3219
3496
|
|
|
3497
|
+
if (!['setup', 'hosts', 'sync-hosts', '-h', '--help', '-v', '--version'].includes(command)) maybeSetupHint();
|
|
3498
|
+
if (command === 'status') maybeNewHostsNotice();
|
|
3499
|
+
|
|
3220
3500
|
switch (command) {
|
|
3221
3501
|
case 'init':
|
|
3222
3502
|
handleInit();
|
|
@@ -3311,6 +3591,12 @@ switch (command) {
|
|
|
3311
3591
|
case 'sync-hosts':
|
|
3312
3592
|
handleSyncHosts();
|
|
3313
3593
|
break;
|
|
3594
|
+
case 'hosts':
|
|
3595
|
+
handleHosts();
|
|
3596
|
+
break;
|
|
3597
|
+
case 'setup':
|
|
3598
|
+
handleSetup();
|
|
3599
|
+
break;
|
|
3314
3600
|
case 'generate-adapter':
|
|
3315
3601
|
handleGenerateAdapter();
|
|
3316
3602
|
break;
|
package/core/grill-policy.md
CHANGED
|
@@ -64,7 +64,7 @@ When the change can touch state, siblings, actors/approvals, or downstream work,
|
|
|
64
64
|
|
|
65
65
|
On `adapter_type: dot`, this is mandatory. On other adapters, skip only when the change cannot touch those four (copy, CSS, comment).
|
|
66
66
|
|
|
67
|
-
Canonical skill: `adapters/
|
|
67
|
+
Canonical skill (adapter-neutral): `adapters/quality/skills/task-impact-inquiry/SKILL.md`. Host copies via `sync-hosts`: `~/.claude/skills/task-impact-inquiry/`, `~/.agents/skills/task-impact-inquiry/` (Codex, Gemini CLI, Grok, Cursor, Copilot, and other shared-folder hosts), `~/.gemini/config/skills/task-impact-inquiry/` (Antigravity). Load that skill if present. If it is missing, still run this section. Do not claim the skill ran.
|
|
68
68
|
|
|
69
69
|
If Stage 1 grill already produced the matrix and the user confirmed it, do not run a second interview.
|
|
70
70
|
|
|
@@ -100,6 +100,17 @@ Native Subagent Invocation: UNAVAILABLE
|
|
|
100
100
|
Review Method: Clean-Slate Artifact Isolation Barrier
|
|
101
101
|
```
|
|
102
102
|
|
|
103
|
+
If native subagent invocation is AVAILABLE but no reviewer subagent returned its own result, the mode line carries the reason, so the header is never contradictory (#57):
|
|
104
|
+
|
|
105
|
+
```text
|
|
106
|
+
Execution Mode: CONTEXT_ISOLATION_ONLY (reason: <why>)
|
|
107
|
+
Independent LLM Execution: NOT PROVEN
|
|
108
|
+
Native Subagent Invocation: AVAILABLE
|
|
109
|
+
Review Method: Clean-Slate Artifact Isolation Barrier
|
|
110
|
+
```
|
|
111
|
+
|
|
112
|
+
REPORT_ONLY (recipe `audit`) policy: when the host can invoke a native subagent, one fresh `devil-advocate` sibling reviews `report.json` before `gate report` (no Judge, no iteration loop); otherwise the same check runs in-session with a reason. `report.json` may carry a `review` block, which `gate report` validates.
|
|
113
|
+
|
|
103
114
|
### Strictly Forbidden Phrases during `CONTEXT_ISOLATION_ONLY`:
|
|
104
115
|
- *"independent agent review"*
|
|
105
116
|
- *"subagent"*
|
package/core/runtime-safety.md
CHANGED
|
@@ -97,8 +97,9 @@ The CLI may update itself, but only under the policy the user chose (`~/.ai-engi
|
|
|
97
97
|
|
|
98
98
|
- **Check:** `run`, `status`, `init`, and `refresh` start `update auto --background` detached at most once every 24 hours (`npm view <pkg> dist-tags`, 3 s timeout). The command itself never waits. Offline or failed checks keep the cached tags and stay silent.
|
|
99
99
|
- **Opt-out:** `AEL_NO_AUTO_UPDATE=1`, `CI=true`, or `update policy off`.
|
|
100
|
-
- **Apply (global installs only):** acquire `~/.ai-engineering-loop/update.lock`, verify the release (registry `repository.url` must match this package, and `dist.integrity` must exist), `npm install -g <pkg>@<target>`, `sync-hosts`, then `refresh --non-destructive` in every project in `~/.ai-engineering-loop/projects.json`. User-edited context files are kept and new templates are written as `*.ael-new`. Projects with an active run are skipped and retried later. Updates are deferred entirely while any UNATTENDED run is active.
|
|
100
|
+
- **Apply (global installs only):** acquire `~/.ai-engineering-loop/update.lock`, verify the release (registry `repository.url` must match this package, and `dist.integrity` must exist), `npm install -g <pkg>@<target>`, `sync-hosts`, then `refresh --non-destructive` in every project in `~/.ai-engineering-loop/projects.json`. User-edited context files are kept and new templates are written as `*.ael-new`. Projects with an active run are skipped and retried later; a human closes a stale run with `run --abandon --reason <text> --by <name>` (agents never abandon a run on their own). Updates are deferred entirely while any UNATTENDED run is active.
|
|
101
101
|
- **Project installs (devDependency or npx):** notify only. `package.json` and lockfiles are never changed automatically. The notice shows current -> latest, 2-4 CHANGELOG highlights, one update command for the detected package manager (npm/pnpm/yarn/bun), and how to silence it (`AEL_NO_UPDATE_NOTICE=1` or `update policy off`). It appears at most once per 24 h per project, on an interactive stderr only, never in CI or `--json` output. Global `notify` and major notices use the same format.
|
|
102
|
+
- **Host setup (no install hooks):** the package defines no `preinstall`/`install`/`postinstall` script, so installing never writes to `$HOME`. `ai-engineering-loop setup` (interactive, or `--yes`) detects agents by config folder or by an executable on `PATH` (stat only, never executed) and saves the selection in `~/.ai-engineering-loop/config.json` (`hosts.enabled`, `hosts.disabled`, `hosts.mode`). `hosts enable|disable <id...>` change it; disabling removes only manifest-tracked, unedited AEL files. Before `setup`, sync writes only to hosts whose config folder exists. Agents never run `setup` or `hosts enable|disable` on their own initiative. Agents installed later are re-detected (stat only): with `hosts.mode: auto` the background `update auto` of a global install enables and syncs them (never a host the user disabled); with `manual` `status` shows one rate-limited notice. Neither writes or prints in CI, in `--json` output, or with `AEL_NO_AUTO_UPDATE`. `hosts doctor` is read-only.
|
|
102
103
|
- **Legacy projects (no `.ael-generated.json` yet):** the first `refresh --non-destructive` overwrites AEL-managed context files directly and deletes obsolete AEL-managed files instead of writing `*.ael-new`. Old copies go to `.ai-engineering-loop/.ael-backup/<timestamp>/` first, and hashes are recorded so later updates use the normal flow. Append-only files (lessons, glossary) and all other user files are never touched. `sync-hosts` likewise removes host files it recorded in its manifest that the package no longer ships (unless edited).
|
|
103
104
|
- **Majors:** `update apply --yes` runs `ai-engineering-loop migrate --from <old>` right after the install. `migrations/<major>.js` steps are idempotent and recorded in `~/.ai-engineering-loop/migrations.json`.
|
|
104
105
|
- **Rollback:** `update rollback --yes` reinstalls the previous version, re-syncs host skills, and restores any legacy-update snapshot taken by that update.
|