40 KiB
lolympus-agent — Research-Driven Improvements (Codename: goofy-plum)
Audience: Codex(또는 다른 AI 에이전트)가 self-contained 하게 작업할 수 있도록 작성됨. 기존
plan.md(Phase 0–5)와evaluation-lively-island.md(Phase 6–9)가 이미 production 인프라(shadow eval / judge / drift / promote / tournament)를 완성한 상태를 전제로 한다.충돌 시 우선순위: 본 문서 >
evaluation-lively-island.md>plan.md. 단 Phase 0–9의 인터페이스/스키마/env 컨벤션은 그대로 따른다.참고 논문 (papers/ 내)
- Paper A — arXiv 2602.12430v3 (Xu & Yan, 2026.02) Agent Skills for LLMs. SKILL.md 명세, progressive disclosure(L1/L2/L3), Tool Search Tool, Programmatic Tool Calling, Skill Trust & Lifecycle Governance(G1–G4 gates · T1–T4 tiers).
- Paper B — arXiv 2502.09054v1 (Zellinger et al., 2025.02) Cost-Saving LLM Cascades with Early Abstention. cheap model이 cascade 초반에 abstain → cost −13%, error −5%, abstention +4.1%.
- Paper C — arXiv 2605.06350v1 (Bouchard, 2026.05) Decision-Theoretic Characterization of LLM Cascades. k>2 cascade는 잘 튜닝된 2-model pairwise envelope 대비 우월하지 않음. Pre-generation diagnostic router가 4/5 benchmarks에서 cascade 우위.
1. Context — 왜 이 작업
문제 (현 production agent의 격차)
- 비용 효율의 구조적 한계 —
OnboardingOrchestrator.run()은derived.tier만 보고 tier0/1/2 LLM 호출. classify confidence가 낮아도 동일 비용. Paper B/C 기준 cheap-model generation 자체를 피할 수 있는 pre-generation 결정 지점이 부재. - Tool 사용의 견고성/확장성 한계 —
tool-loop.ts는maxHops=2고정, retry/error 분류 없음. tool 1개 실패 시 throw로 전체 흐름 정지. 신규 skill 도입 시 도구 prompt bloat 폭증 구조. - Skill 추상화 부재 —
PromptBuilder.buildSystemPrompt()가 persona/policy/state/examples/context를 매 요청 1회 직렬화(~1800 tokens). 신규 skill 추가가 단기 결정됨에 따라 SKILL.md progressive disclosure 도입 필수. - 자동 propose의 안전망 부재 —
eval/loop/propose.ts는 LLM이 생성한 prompt variant를 직접 file에 기록. injection / unintended behavior 차단 게이트 없음. Paper A: community skill 26.1%에 취약점.
의도한 결과 (3개월 horizon)
- 평균 응답 비용 −15~20%, p50 latency −10~15%
- LLM judge overall regression 0 (또는 +0.3 이상)
- 신규 skill 추가 시간을 monolithic 대비 5x 단축할 수 있는 SKILL.md 인프라
- 자동 propose 변형이 G1–G4 gate를 거쳐 prod에 안전 승급되는 폐쇄 루프
비목표
- 코드 실행 sandbox 형태의 Programmatic Tool Calling (DAG 형태로 한정)
- 외부 embedding 서비스 의존 (로컬 Qwen-embed 만 사용; 미정인 동안 v3 router 보류)
- 단일 회사 외부 trust tier T4(vendor-certified) 자동화 — 인간 검토 유지
2. 사용자 결정 (반영됨)
| 항목 | 결정 |
|---|---|
| Phase α 우선 가치 | 균형 — Early Abstention + Tool Error Recovery + Programmatic 시범 |
| Abstention 정책 | Hybrid (대화 위치 기반) — 초반 turn은 짧은 확인 질문, 반복/혼란 시그널은 [NO_RESPONSE] |
| Skill 확장 | 단기 내 추가 예정 → SKILL.md(#1) Phase β로, Governance(#7) Phase γ 앞부분으로 |
3. Relationship to Existing Plans (변경 영향)
그대로 사용 (변경 없음)
src/eval/{capture,anonymize,runner,types,shadow-bus,shadow-worker,aggregator,archive,drift,promote}.tssrc/eval/judges/{rule-judge,llm-judge}.tssrc/eval/loop/{tournament,elo}.tssrc/eval/review/{queue,label-archive,calibration}.tssrc/eval/golden/*.jsonl(스키마 호환 유지)- Redis/Garnet 키, NATS subjects, env 컨벤션
- 익명화(
anonymize.ts) — 신규 user-memory 쓰기에 강제 적용
확장 / 변경
src/skills/onboarding/runtime/{orchestrator,tool-loop,critic,trace}.ts— 행동 분기 추가, 시그니처 일부 확장src/skills/onboarding/state/{derive,onboarding-state,blackboard}.ts—routingDecision,abstentionScore, long-term key 추가src/skills/onboarding/prompts/{builder,layers,examples}.ts— SKILL.md mode 분기, retriever 호출src/llm/routing.ts—selectModelsForProvider·selectProviderOrder시그니처에routingDecision옵션 추가src/eval/judges/process-judge.ts— PRM forward score 통합src/eval/loop/propose.ts— T1 자동 진입 + gate 호출src/config.ts— 신규 env 항목 (Section 4 참조)
신규 파일
| 경로 | 역할 | Phase |
|---|---|---|
src/skills/onboarding/runtime/abstention.ts |
Hybrid early abstention 결정 | α |
src/skills/onboarding/runtime/tool-plan-runner.ts |
DAG 기반 programmatic tool calling | α |
src/skills/onboarding/runtime/router.ts |
Pre-gen diagnostic router | β |
src/skills/onboarding/runtime/runtime-flag-manifest.ts |
tournament 가능한 runtime flag variant | α (시작) / β (확장) |
src/skills/onboarding/SKILL.md |
L1 metadata + L2 body | β |
src/skills/onboarding/resources/*.md |
L3 resources (nickname-format, heart-rule, discord-format, age-policy) | β |
src/skills/skill-loader.ts |
SkillRegistry — 모든 SKILL.md 인덱스 | β |
src/skills/onboarding/state/user-memory.ts |
UserMemoryStore (long-term) | β |
src/skills/onboarding/prompts/example-store.ts |
ExampleRetriever | β |
src/eval/judges/prm.ts |
Process Reward Model (v1 Ridge) | β |
src/eval/judges/attribution.ts |
outcome → trace step root cause | β |
src/eval/governance/gates.ts |
G1–G4 gates | γ |
src/eval/governance/trust.ts |
T1–T4 promotion/demotion | γ |
src/tools/tool-search.ts |
search_tools meta tool | γ |
신규 의존
- 없음. 모든 신규 기능은 기존 의존(Garnet, MongoDB driver, Zod, LlmClient, Pino, OpenTelemetry)로 구현 가능.
- (v3 router·example-store v2 시점에) 로컬 Qwen-embed가 필요 — 그 시점 별도 결정 (현 단계 보류).
4. 신규 환경 변수 / 설정
src/config.ts에 추가. 모든 신기능은 default OFF로 시작 → shadow eval 후 ON.
# Phase α
AGENT_ABSTENTION_ENABLED=false
AGENT_ABSTENTION_THRESHOLD=0.35 # ξ — [NO_RESPONSE] 임계
AGENT_ABSTENTION_CLARIFY_THRESHOLD=0.45 # φ_clarify — clarify 질문 임계
AGENT_TOOL_RETRY_MAX=1
AGENT_TOOL_LOOP_MAX_HOPS=3 # 기존 2 → 3 (precondition retry 여유)
AGENT_PROGRAMMATIC_TOOL_CALLING=false # variant manifest 노출도 가능
# Phase β
AGENT_SKILL_MD_MODE=false # SKILL.md L1/L2/L3 disclosure
AGENT_ROUTER_MODE=static # static | learned-v1 | learned-v2
AGENT_ROUTER_WEIGHTS_PATH=
AGENT_USER_MEMORY_ENABLED=false
AGENT_USER_MEMORY_TTL_DAYS=30
EVAL_PRM_ENABLED=false
EVAL_PRM_WEIGHTS_PATH=
# Phase γ
AGENT_TRUST_MIN_TIER=T2 # 미달 variant는 control fallback
AGENT_CRITIC_MAX_ITERATIONS=2
AGENT_TOOL_SEARCH_ENABLED=false
5. Phase α (Week 1-2) — 즉시 효과·낮은 위험
TASK-α-2-01 · Hybrid Early Abstention
파일: src/skills/onboarding/runtime/abstention.ts (신규), src/skills/onboarding/runtime/orchestrator.ts, src/skills/onboarding/state/onboarding-state.ts, src/config.ts, src/skills/onboarding/onboarding-templates.ts
시그니처:
// src/skills/onboarding/runtime/abstention.ts
export type AbstentionAction =
| { kind: "proceed" }
| { kind: "clarify"; templateKey: "abstain-clarify-default" | string }
| { kind: "no-response" };
export interface AbstentionDecision {
action: AbstentionAction;
reason:
| "low-confidence-early-turn"
| "low-confidence-repeated"
| "off-topic-strong"
| "flagged-confused"
| null;
score: number; // 0..1
features: {
classifyConfidence: number;
intentKind: Intent["kind"];
turnIndex: number;
sameIntentStreak: number;
flaggedIssues: number;
};
}
export function decideEarlyAbstention(input: {
classified: ClassifiedIntent;
derived: DerivedState;
blackboard: Blackboard | null;
request: AgentRequest;
thresholds: {
abstain: number; // AGENT_ABSTENTION_THRESHOLD
clarify: number; // AGENT_ABSTENTION_CLARIFY_THRESHOLD
};
}): AbstentionDecision;
룰 (Hybrid):
intent.kind === "off-topic-to-other"&&confidence > 0.9→no-response, reason=off-topic-strong.flaggedIssues.length > 0||sameIntentStreak >= 3(blackboard.lastIntent 비교) →no-responseif confidence < abstain, reason=flagged-confusedorlow-confidence-repeated.turnIndex <= 1(greetedAt 미존재) &&confidence < clarify&& intent ∈ {unknown, chitchat} →clarify(templateKey=abstain-clarify-default= "조금 더 구체적으로 알려주시면 도와드릴게요!"), reason=low-confidence-early-turn.- 그 외 →
proceed.
orchestrator 통합 (runtime/orchestrator.ts:run() 진입부 직후):
if (getAbstentionEnabled()) {
const decision = decideEarlyAbstention({
classified, derived, blackboard, request,
thresholds: { abstain: getAbstentionThreshold(), clarify: getAbstentionClarifyThreshold() },
});
derived.abstentionScore = decision.score;
traceCollector?.startStep("classify").finish({
confidence: classified.confidence,
intent: classified.intent.kind,
source: classified.source,
abstention: decision, // process-trace에 기록
});
if (decision.action.kind === "no-response") {
return { text: "[NO_RESPONSE]", validation: { status: "pass", reasons: ["early-abstention"] } };
}
if (decision.action.kind === "clarify") {
return { text: getOnboardingTemplate(decision.action.templateKey), validation: { status: "pass", reasons: ["early-abstention-clarify"] } };
}
}
템플릿 추가 (onboarding-templates.ts):
"abstain-clarify-default": "조금 더 구체적으로 알려주시면 도와드릴게요!",
Trace 확장 (runtime/trace.ts):
ProcessTraceStep<classify>.output에abstention?: AbstentionDecision필드 추가 (옵셔널, 기존 호환).
Eval 연동:
src/eval/aggregator.ts: slice에abstentionReason,abstentionAction차원 추가.src/eval/drift.ts: abstention rate (action≠proceed 비율)를 metric으로 등록.- 통과 기준: 7일 shadow에서 cost −10% 이상, false-abstention < 2%, quality regression < 0.05.
Acceptance:
pnpm test src/skills/onboarding/runtime/abstention.test.ts(신규 unit test 5+ case: off-topic / early-low-conf / repeated / proceed / boundary)- env OFF에서 기존 behavior 100% 일치 (snapshot 비교)
- env ON 24h shadow에서 abstention rate 0.05~0.25 사이 (밴드 밖이면 임계 재튜닝)
TASK-α-8a-01 · Structured Tool Error Recovery
파일: src/skills/onboarding/runtime/tool-loop.ts, src/tools/tool.ts, src/skills/onboarding/runtime/trace.ts, src/config.ts
타입 확장 (src/tools/tool.ts):
export type ToolErrorKind = "args-invalid" | "transient" | "precondition-failed" | "permanent";
export interface ToolExecutionFailure {
kind: ToolErrorKind;
toolName: string;
message: string;
retryable: boolean;
details?: unknown;
}
tool-loop.ts 변경 요지 (현 :113-137 영역):
- zod parse 실패 →
args-invalid→ 1 hop 자기교정 (assistant tool_call → user "이전 args가 schema 어긋남" 안내). 이때 hop budget을 consume 하되 별도 카운터selfCorrectionHops추적해 무한 루프 차단(max 1). - fetch/timeout 등 network error 감지 (
AbortError/FetchError/timeout) →transient→ 200ms backoff → 1회 retry (AGENT_TOOL_RETRY_MAX). hop budget 사용 안 함. - tool 내부 throw →
permanent→{ ok: false, errorKind: "permanent", message }를 tool 결과로 message에 push, hop 계속. - precondition 검증 실패 (예: session 없음) → tool이
result.ok=false로 정상 반환하도록 builtins 수정 → tool-loop에서precondition-failed로 분류, hop budget consume 안 함.
maxHops: 기본 getAgentToolLoopMaxHops() (default 3). precondition-failed로 인한 추가 hop은 +1 까지만 허용.
Trace 확장:
toolCalls?: Array<ProcessTraceStep<{
name: string;
args: unknown;
ok: boolean;
resultSummary: string;
errorKind?: ToolErrorKind; // 신규
retried?: boolean; // 신규
}>>;
재사용:
parseToolCalls(src/tools/parser.ts) — 변경 없음- 기존
traceStep?.finish({ ok: false, ... })경로
Acceptance:
src/skills/onboarding/runtime/tool-loop.test.ts(신규): mock tool throw / network timeout / zod-invalid / precondition-failed 각 케이스- tool-heavy slice(nickname-rejected, discord-verify-question) shadow eval에서 success rate +3% 이상
- p95 latency 변화 < +200ms
TASK-α-4a-01 · Programmatic Tool Calling 시범 (선언적 DAG)
파일: src/skills/onboarding/runtime/tool-plan-runner.ts (신규), src/skills/onboarding/runtime/planner-schema.ts, src/skills/onboarding/runtime/orchestrator.ts
스키마 확장 (planner-schema.ts):
export const toolPlanStepSchema = z.object({
id: z.string().min(1),
tool: z.string().min(1),
argsFromContext: z.record(z.string(), z.string()).optional(), // "$request.nickname"
argsLiteral: z.record(z.string(), z.unknown()).optional(),
dependsOn: z.array(z.string()).optional(),
on: z.object({ error: z.enum(["abort", "fallback-llm"]).optional() }).optional(),
});
export const toolPlanSchema = z.object({
steps: z.array(toolPlanStepSchema).max(4), // 안전 한계
composeWith: z.enum(["all-then-llm", "first-then-llm"]),
});
export const plannerOutputSchema = z.object({
approach: z.string(),
mustInclude: z.array(z.string()).optional(),
mustAvoid: z.array(z.string()).optional(),
suggestedTools: z.array(z.string()).optional(),
targetLengthChars: z.number().optional(),
toolPlan: toolPlanSchema.optional(), // NEW
});
runner (tool-plan-runner.ts):
export interface ToolPlanResult {
results: Record<string /* step id */, unknown>;
failures: ToolExecutionFailure[];
durationMs: number;
}
export async function runToolPlan(
plan: ToolPlan,
tools: readonly Tool<unknown, unknown>[],
ctx: ToolContext,
request: AgentRequest,
): Promise<ToolPlanResult>;
- DAG 정렬 (
dependsOn) 후Promise.all로 동일 레벨 동시 실행 (단 동일 NATS RPC 큐 부담 고려해Promise.all최대 2 병렬). $request.nickname,$request.userId,$session.openQuestions[0]등 path 표현식 resolve.- 실패 시
on.error === "fallback-llm"이면 tool-loop으로 fallback (free-form),abort이면 즉시 종료.
orchestrator 통합 (run() planner 출력 처리 부분):
- planner output에
toolPlan있고getProgrammaticToolCalling()ON이면runToolPlan먼저 → 결과를 message[]에 tool messages로 push → responder 1회 호출. - 없으면 기존
runWithTools경로.
variant 노출:
src/eval/proposals/planner-with-toolplan.json신규 — planner system layer에"가능하면 toolPlan 필드를 채워주세요"append.- Tournament로 control vs toolplan 비교. 통과 기준 ELO +20, latency p50 −150ms 이상.
Acceptance:
src/skills/onboarding/runtime/tool-plan-runner.test.ts(신규): 정상 DAG / 의존성 누락 / step 실패 fallback 케이스- process-judge 확장:
toolPlan.steps.lengthvsactualToolCalls.lengthconsistency check (src/eval/judges/process-judge.ts)
TASK-α-FLAG-01 · RuntimeFlagManifest 인프라
파일: src/skills/onboarding/runtime/runtime-flag-manifest.ts (신규), src/skills/onboarding/runtime/orchestrator.ts, src/eval/loop/tournament.ts
타입:
export interface RuntimeFlagManifest {
variantId: string; // e.g. "abstention-on", "router-v1"
description: string;
flags: Partial<{
abstentionEnabled: boolean;
abstentionThreshold: number;
abstentionClarifyThreshold: number;
programmaticToolCalling: boolean;
toolLoopMaxHops: number;
routerMode: "static" | "learned-v1" | "learned-v2";
userMemoryEnabled: boolean;
criticMaxIterations: number;
}>;
}
export function loadRuntimeFlagManifest(variantId: string): RuntimeFlagManifest | null;
export function resolveRuntimeFlags(
base: ResolvedRuntimeFlags,
variant?: RuntimeFlagManifest,
): ResolvedRuntimeFlags;
효과: PromptVariantManifest(prompt만)와 별도로 행동 변화 variant를 tournament로 비교 가능. tournament.ts에서 loadVariantBundle(variantId)가 prompt+runtime 둘 다 로드하도록 확장.
Acceptance:
runtime-flag-manifest.test.ts: variant 로드 / merge / 잘못된 키 거부- tournament가 동일 골든셋에 대해 control / abstention-on / toolplan-on 3종 ELO 산출 가능
Phase α Exit 게이트
전부 만족 시 Phase β 진입.
| 메트릭 | 측정 위치 | 목표 |
|---|---|---|
| Cost (shadow 7일 평균) | eval/aggregator.ts slice별 latencySumMs * tier price |
−10% vs control |
| LLM judge overall | aggregator.llmOverallAvg |
≥ baseline − 0.05 |
| Abstention false-positive | review/queue.ts 라벨 vs decision |
< 2% |
| Tool failure → soft-template | validator.ts reject 사례 비율 |
−30% |
| Regression alerts | drift.ts significant count |
0 |
6. Phase β (Week 3-5) — Skill 확장 대비 + 평가 정밀화
TASK-β-1-01 · SKILL.md L1/L2/L3 Progressive Disclosure
핵심 가정: 단기 내 추가 skill(discord-help / community-rules / identity-room 중 1+) 도입 결정. SKILL.md 인프라가 없으면 prompt 폭증.
신규 파일:
src/skills/onboarding/SKILL.md— frontmatter + bodysrc/skills/onboarding/resources/{nickname-format,heart-rule,discord-format,age-policy}.mdsrc/skills/skill-loader.ts— SkillRegistry
SKILL.md frontmatter 스키마:
---
name: onboarding
description: 카카오톡 오픈채팅 새내기 등록을 도와주는 친절한 선배 역할.
version: 1
intents: [greeting, acknowledgement, status-question, help-how, help-where, nickname-question, photo-question, discord-question, discord-verify-question, age-question]
triggers: [user-message, new-member, nickname-invalid, photo-failed, progress-update, timeout-warning, completed]
tools: [validate-nickname, check-onboarding-progress, lookup-community-rule, get-room-context, verify-discord-nickname]
resources:
nickname-format: ./resources/nickname-format.md
heart-rule: ./resources/heart-rule.md
discord-format: ./resources/discord-format.md
age-policy: ./resources/age-policy.md
trustTier: T3
---
SkillRegistry:
// src/skills/skill-loader.ts
export interface SkillManifest {
name: string;
description: string;
version: number;
intents: string[];
triggers: string[];
tools: string[];
resources: Record<string, string>; // key → relative path
trustTier: TrustTier;
bodyPath: string; // SKILL.md absolute path
}
export class SkillRegistry {
static load(rootDir: string): SkillRegistry; // 모든 SKILL.md frontmatter 메모리 로드
list(): SkillManifest[];
get(name: string): SkillManifest | null;
loadBody(name: string): Promise<string>; // L2 lazy load
loadResource(name: string, key: string): Promise<string>; // L3 lazy load
}
PromptBuilder 재설계 (prompts/builder.ts):
- env
AGENT_SKILL_MD_MODE=true일 때:- system 1차 메시지: 모든 skill manifest의
description만 ~30 tokens/skill 직렬화 (L1). - 매칭된 skill의 body(
loadBody)를 system 2차 메시지로 (L2). - planner output에
requestResources: ["nickname-format"]있으면 responder 호출 직전에 해당 resource MD를 system 3차 메시지로 inject (L3).
- system 1차 메시지: 모든 skill manifest의
- 기존 layer 함수(
personaLayer,policyLayer, ...)는 그대로 두되policyLayer만 intent-conditional layer로 분리되어 L2 단계에서 inject. (점진 마이그레이션)
planner-schema 확장:
requestResources: z.array(z.string()).optional(),
마이그레이션:
- Step 1: SKILL.md를 mirror로 생성. 기존 prompt 결과와 token 차이 측정.
- Step 2: feature flag ON 후 24h shadow eval. quality regression < 0.05 확인.
- Step 3: L3 resource를 nickname-related intent부터 적용 (가장 적은 intent → 점진).
Acceptance:
- SkillRegistry unit test: 잘못된 frontmatter / 빈 resources / 중복 name
- snapshot test: system prompt token count baseline ↔ SKILL.md mode (≤ −40%)
- 30일 long-soak shadow: regression < 0.05
- dry-run: 새 skill 추가가 SKILL.md 1개 + registry 1줄로 동작
TASK-β-5-01 · Process Reward Model (PRM) + Bidirectional Attribution
신규 파일:
src/eval/judges/prm.ts— feature extractor + Ridge regression (사용 라이브러리: 직접 구현, dependencies 최소)src/eval/judges/attribution.ts— outcome → trace step root cause
Feature vector (Section 5 #5 PRM 참조):
export interface PrmFeatureVector {
classifyConfidence: number; classifySource: 0|1;
planApproach: string; // categorical hash
toolCallCount: number; toolErrorCount: number;
criticDecision: 0|1; validateStatus: 0|1|2; tier: 0|1|2;
abstentionScore: number; // Phase α에서 추가됨
}
Forward model (v1): Ridge regression on label-archive humanScore vs features. weights는 EVAL_PRM_WEIGHTS_PATH(JSON)에 저장.
Backward attribution:
- 각 entry에 대해 trace step별 counterfactual delta (step output 평균값으로 swap 시 PRM score 변화) →
attributedScoreLoss. - 결과 collection:
eval_process_attribution(MongoDB).
process-judge 통합 (eval/judges/process-judge.ts):
// 기존 ProcessJudgeResult에 추가
prmScore?: number; // PRM forward
prmFeatures?: PrmFeatureVector;
rootCauseStep?: ProcessTraceStepName;
rootCauseScoreLoss?: number;
Background job 추가 (workers/background.ts):
- 새 tick:
EVAL_ATTRIBUTION_TICK_MS(default 6h). 최근 archive entries에 대해 attribution 계산 후 collection 적재. - 라벨 수가 N(env
EVAL_PRM_MIN_LABELS, default 500) 미만이면 forward 모델은 "explanation only" 모드(score 출력 X, root cause만).
propose 입력 확장 (eval/loop/propose.ts):
buildProposalInputFromArchive에rootCauses: RootCauseReport[]필드 추가. LLM이 root cause 분포를 보고 더 타겟팅된 prompt diff 생성.
Acceptance:
prm.test.ts: feature vector 안정성, weight 로드/저장attribution.test.ts: synthetic trace에서 root cause step 식별 정확도 ≥ 0.8- tournament: PRM-augmented variant가 ELO +30 차이를 식별 가능 (synthetic label)
TASK-β-3-01 · Pre-Generation Diagnostic Router (v1 룰)
신규 파일: src/skills/onboarding/runtime/router.ts
Decision space:
export type RoutingDecision =
| { kind: "abstain"; reason: string } // = TASK-α-2 abstention
| { kind: "template-only"; templateKey: string }
| { kind: "tier0-only" }
| { kind: "tier1-direct"; skipPlanner: boolean }
| { kind: "tier1-with-tools" }
| { kind: "tier2-full" };
export interface DiagnosticRouter {
decide(input: RouterInput): RoutingDecision;
}
v1 rules (router.ts 초안):
intent.kind === "off-topic-to-other"&& conf > 0.9 →abstainintent.kind === "greeting"&& conf > 0.95 && no prior greeting →template-onlyintent.kind === "acknowledgement"&& conf > 0.9 →template-onlyphase.kind === "nickname-rejected"→tier1-with-toolsphase.kind === "completed"&& intent ≠ status-question →template-onlyphase.kind ∈ {fresh, nickname-pending, photo-pending, both-pending}&& intent ∈ {nickname-question, photo-question, help-how} →tier1-with-tools- 그 외 →
tier0-only
derive.ts 통합: deriveState() 끝에서 router.decide(...) 호출, 결과를 DerivedState.routingDecision에 저장.
llm/routing.ts 확장:
selectModelsForProvider(provider, options, llamaFallbackModel)시그니처에routingDecision?: RoutingDecision추가.- llama 분기(
routing.ts:24-25)는routingDecision.kind가tier1-with-tools이상이면 fallback 모델 대신 explicit fallback 사용; 현 단일 모델 환경에서는 effective change 0 (multi-llama 환경 대비 hook 추가). selectProviderOrder(endpoints, options)도routingDecision을 받아 cliproxy/openrouter/llama 우선순위를 동적 조정 (예:template-onlydecision이면 llama only).
Acceptance:
router.test.ts: 30+ 케이스로 v1 룰 커버- Shadow A/B 24h: cost −8%, quality flat (LLM judge overall delta < 0.05)
routingDecision을ShadowEvalEnvelope에 기록 (shadow-bus.ts) — aggregator slice 추가
TASK-β-3-02 · Router v2 학습
의존: TASK-β-5-01 (PRM features). 데이터 ≥ 5000 archive entries.
파일: src/skills/onboarding/runtime/router.ts 확장, scripts/train-router.ts (신규 script)
모델: lightweight logistic regression (multi-class: 6 decision kinds). 입력은 PRM extractFeatures + 사용자 메시지 길이/언어 식별.
학습 데이터 라벨링:
- archive에서 actual derived.tier · validation status · llmJudge.overall 기반 ground truth tier 추정:
- llmJudge.overall ≥ 4 + tier=0 →
tier0-only충분 - llmJudge.overall < 3 + tier=0 →
tier1-direct필요했음 (under-routed) - llmJudge.overall < 3 + tier=1 →
tier2-full필요했음
- llmJudge.overall ≥ 4 + tier=0 →
- 5% exploration noise (
epsilon=0.05) 도입 — 무작위로 1 tier ↑
Acceptance:
train-router.ts가 weights JSON 생성 (AGENT_ROUTER_WEIGHTS_PATH)- Tournament:
router-v1vsrouter-v2비교 — v2 cost −5% 추가 절감
TASK-β-6-01 · Persistent User Memory + Adaptive Few-shot
신규 파일:
src/skills/onboarding/state/user-memory.tssrc/skills/onboarding/prompts/example-store.ts
Garnet 키: lolympus-agent:user:profile:{anonHash(userId)} (TTL AGENT_USER_MEMORY_TTL_DAYS × 86400).
UserMemoryStore:
export interface UserProfile {
userId: string; // anonHash
firstSeenAt: string; lastSeenAt: string;
totalTurns: number;
preferredStyle: "concise" | "warm" | "neutral";
pastIntents: { kind: Intent["kind"]; count: number }[];
flaggedIssues: string[];
completedOnboardingAt?: string;
}
export class UserMemoryStore {
constructor(private garnet: GarnetClient, private salt: string) {}
async load(userId: string): Promise<UserProfile | null>;
async observeTurn(userId: string, derived: DerivedState, response: AgentResponse): Promise<UserProfile>;
}
PII 강제:
- 모든 string write에
anonymizeText(src/eval/anonymize.ts) 적용. unit test로 nickname/이메일/번호 미저장 검증. - userId 자체는
anonymizeIdentifier("user", userId)해시 사용.
ExampleRetriever (prompts/example-store.ts):
- v1: intent + phase로 필터링한 golden examples 중 user.preferredStyle과 매칭되는 최대 3개 반환.
- v2 (embedding): 보류 — 로컬 Qwen-embed 도입 결정 후.
examplesLayer 시그니처 확장 (prompts/examples.ts):
export function examplesLayer(intent: Intent, retriever: ExampleRetriever, query: ExampleQuery): string;
variant 노출: runtime-flag-manifest userMemoryEnabled: true.
Acceptance:
user-memory.test.ts: PII redaction 강제, observeTurn idempotent- 같은 user(anon) 재방문 시 LLM judge friendliness 평균 +0.2 (shadow eval)
Phase β Exit 게이트
| 메트릭 | 목표 |
|---|---|
| system prompt token | baseline 대비 −40% (SKILL.md mode ON) |
| 누적 cost (β 종료) | α 대비 추가 −8% |
| PRM attribution AUC | ≥ 0.75 (라벨 1000+) |
| 신규 skill dry-run | SKILL.md 1개 + registry 1줄로 동작 확인 |
7. Phase γ (Week 6+) — Governance + Reflection
TASK-γ-7-01 · Skill Trust & Lifecycle Governance (G1–G4 · T1–T4)
신규 파일: src/eval/governance/gates.ts, src/eval/governance/trust.ts
Gates 시그니처:
export interface GateResult { passed: boolean; reason?: string; evidence?: unknown; }
export function runG1Static(manifest: PromptVariantManifest | RuntimeFlagManifest): GateResult;
// URL allowlist (validator.ts:8-11 재사용), layer 길이 제한, persona override 차단, 위험 키워드
export async function runG2Semantic(manifest, control, llmClient: LlmClient): Promise<GateResult>;
// LLM이 변형의 핵심 행동 변화 의도 분류 — "변경이 의도된 범위 내인가?"
export async function runG3Sandbox(manifest, evalRunner, goldenSetPath, n: number = 200): Promise<GateResult>;
// shadow run on N=200 cases, validator reject rate / llm overall 변화 측정
export function runG4Permission(manifest, toolRegistry: ToolRegistry, skillRegistry: SkillRegistry): GateResult;
// required tools/resources가 registry에 등록되어 있는지
Trust Tiers (trust.ts):
export type TrustTier = "T1-unvetted" | "T2-community" | "T3-org-vetted" | "T4-certified";
// Promotion 규칙
// T1 → T2: G1 + G4 통과 (자동)
// T2 → T3: G3 sandbox + tournament 100+ matches + ELO > control + 30 + drift OK 7일 (자동)
// T3 → T4: human review label-archive 평균 ≥ 4.0 + reviewer 2명 이상 동의 (semi-auto)
//
// Demotion: drift significant 또는 G3 sandbox regression → 즉시 T1로 강등 + alert
export async function evaluatePromotion(variantId: string, services: GovernanceServices): Promise<TrustTier>;
export async function applyDemotion(variantId: string, reason: string): Promise<void>;
propose.ts 통합 (eval/loop/propose.ts):
- 새 proposal은 항상 T1로 저장.
writeVariantManifest호출 직후runG1Static+runG4Permission실행 → 통과 시 T2로 자동 승급.
런타임 강제 (src/skills/onboarding/skill.ts):
getPromptVariantId()로드 시trust.getActiveTier(variantId)확인.AGENT_TRUST_MIN_TIER(default T2) 미달이면 control fallback.
Background job:
workers/background.ts에EVAL_PROMOTION_TICK_MS(default 30m). 모든 T1/T2 variant 대상 promotion 평가 실행.
Acceptance:
gates.test.ts: 의도적 prompt injection variant가 G1에서 차단- T1→T2 자동 승급, T2→T3 tournament 게이트 통과 시 자동 승급 시뮬레이션
- Drift significant 발생 시 즉시 강등 통합 테스트
TASK-γ-4b-01 · Tool Search Tool
신규 파일: src/tools/tool-search.ts. src/tools/registry.ts에 searchByDescription(query, k) 추가.
export const toolSearchTool: Tool<{ query: string; topK?: number }, { tools: Array<{ name: string; description: string }> }> = {
name: "search_tools",
description: "도구 카탈로그에서 가장 적합한 도구 후보를 검색합니다.",
argsSchema: z.object({ query: z.string().min(2), topK: z.number().int().min(1).max(5).default(3) }),
resultSchema: z.object({ tools: z.array(z.object({ name: z.string(), description: z.string() })) }),
async execute(args, ctx) { /* registry.searchByDescription */ },
};
registry 변경 (tools/registry.ts):
searchByDescription(query, k): 단순 keyword overlap → BM25 (lite, 의존 추가 없이 inline 구현).- 도구 등록 시 description 인덱싱.
orchestrator 통합 (runtime/orchestrator.ts):
- skill manifest의
tools배열이 N 초과(envAGENT_TOOL_SEARCH_THRESHOLD, default 10)일 때만search_tools만 노출 + planner가requestTools: [...]통해 expand.
Acceptance:
- 20 tool mock registry에서 nickname 관련 query → validate-nickname top-1 회수
- 도구 5개 미만 환경에서는 effective change 0 (기존 path)
TASK-γ-8b-01 · Multi-step Self-Critic (Reflection)
파일: src/skills/onboarding/runtime/critic.ts, src/skills/onboarding/runtime/trace.ts
시그니처 확장:
export interface CritiqueResultV2 {
changed: boolean;
text: string;
iterations: number;
abstained: boolean;
rubricScores: { accuracy: number; scope: number; tone: number };
prmScoreBefore?: number;
prmScoreAfter?: number;
}
class OnboardingCritic {
async critiqueIterative(
draft: string, derived: DerivedState, messages: ChatMessage[],
options: { maxIters: number; deadlineMs: number; prm?: PrmModel },
): Promise<CritiqueResultV2>;
}
알고리즘:
- PRM forward score →
prmScoreBefore. score <criticRubricFloor&& tier == 2 → 반복. - rubric 기반 prompt로 LLM 호출 → 점수 + 개선안.
revise()1회. 다시 PRM.prmScoreAfter - prmScoreBefore < δ면 abort.iterations <= maxIters(default 2). 결과 trace에 기록.- PRM이
abstainFloor미만이면[NO_RESPONSE]반환 (Paper B 통합).
Tier 적용: tier 2만. tier 0/1은 기존 1-shot critic 유지.
deadline 강제: 매 iteration 시작 시 Date.now() < deadlineMs 체크. 초과면 즉시 반환.
Acceptance:
- tier 2 slice에서 LLM judge accuracy +0.3
- critic 평균 iterations 2 이하, p99 latency +400ms 이하
Phase γ Exit 게이트
| 메트릭 | 목표 |
|---|---|
| 자동 propose-deploy 주기 | weekly 1+ variant T2 자동 승급 |
| Tier2 PRM rubric score | +0.3 이상 |
| 누적 cost (γ 종료) | baseline 대비 −15~20% |
8. 평가 시스템 통합 (변경별)
| 변경 | shadow-bus | aggregator slice | tournament | drift | promote |
|---|---|---|---|---|---|
| #2 Abstention | decision.reason/action |
abstention rate | — | abstention drift | abstention-mismatch |
| #8a Tool Recovery | errorKind in trace |
tool failure rate | — | tool error drift | tool-failure |
| #4a Programmatic | toolPlan |
tool success rate | plan-mode vs free | tool round-trip | tool-failure |
| #1 SKILL.md | disclosureMode |
system token usage | L1L2L3 vs flat | — | regression |
| #5 PRM | trace 그대로 | PRM score column | PRM-aug vs raw | PRM score drift | rule-prm-disagree |
| #3 Router | routingDecision |
tier mix | router-on vs off | tier-mix drift | wrong-router |
| #6 Memory | userProfileLoadedAt |
재방문 user slice | with-mem vs without | memory hit | low-retention |
| #7 Governance | variantTrustTier |
tier별 quality | 승급 게이트 | demotion alert | T1 강등 |
| #8b Self-Critic | iterations, rubricScores |
tier2 cost | iter-1 vs iter-2 | critic loop drift | critic-loop-failure |
9. 검증 — End-to-End
개발 중
# 1. unit + integration test
pnpm test
# 2. golden set 회귀
pnpm run eval:run --input src/eval/golden/onboarding.jsonl
# ruleJudge.score, llmJudge.overall regression < 0.05 확인
# 3. tournament (variant 등록 후)
pnpm run eval:tournament --variants control,<new-variant>
# ELO 차이 측정
# 4. drift (5분 tick)
pnpm run eval:report --slice tier=0
운영 검증 (각 Phase 종료 후 1주)
- Aggregator dashboard: slice별 quality/cost/latency 추세
- Drift 알람 false-positive 0 (변경 의도와 일치)
- Promote candidates에 의미 있는 root cause 포함
Rollback
- 모든 신규 행동은 env flag 또는
RuntimeFlagManifest뒤에 배치 → 즉시 OFF - shadow eval 단계에서 production 영향 0 (counterfactual 기록만)
10. 핵심 수정 파일 — 진입점 (Codex가 첫 PR에서 손댈 파일)
Phase α 첫 PR (TASK-α-2-01 + TASK-α-FLAG-01):
src/skills/onboarding/runtime/abstention.ts(신규)src/skills/onboarding/runtime/orchestrator.ts— abstention gate 진입src/skills/onboarding/state/onboarding-state.ts—DerivedState.abstentionScore?추가src/skills/onboarding/onboarding-templates.ts— clarify 템플릿src/skills/onboarding/runtime/trace.ts— classify step에abstention?필드src/skills/onboarding/runtime/runtime-flag-manifest.ts(신규)src/config.ts—AGENT_ABSTENTION_*env 추가src/skills/onboarding/runtime/abstention.test.ts(신규)
Phase α 두 번째 PR (TASK-α-8a-01):
src/tools/tool.ts—ToolErrorKind,ToolExecutionFailuresrc/skills/onboarding/runtime/tool-loop.ts— retry/error 분류src/skills/onboarding/runtime/trace.ts— toolCalls output에errorKind?추가src/skills/onboarding/runtime/tool-loop.test.ts(신규)
Phase α 세 번째 PR (TASK-α-4a-01):
src/skills/onboarding/runtime/planner-schema.ts—toolPlan?필드src/skills/onboarding/runtime/tool-plan-runner.ts(신규)src/skills/onboarding/runtime/orchestrator.ts— toolPlan 우선 분기src/eval/proposals/planner-with-toolplan.json(신규 variant)src/eval/judges/process-judge.ts— toolPlan consistency checksrc/skills/onboarding/runtime/tool-plan-runner.test.ts(신규)
Phase β 핵심:
src/skills/skill-loader.ts,src/skills/onboarding/SKILL.md,src/skills/onboarding/resources/*.mdsrc/skills/onboarding/prompts/builder.ts— SKILL.md modesrc/eval/judges/prm.ts,src/eval/judges/attribution.ts,src/eval/judges/process-judge.tssrc/skills/onboarding/runtime/router.ts,src/skills/onboarding/state/derive.ts,src/llm/routing.tssrc/skills/onboarding/state/user-memory.ts,src/skills/onboarding/prompts/example-store.ts
Phase γ:
src/eval/governance/{gates,trust}.tssrc/eval/loop/propose.ts(T1 entry)src/tools/{tool-search.ts,registry.ts}src/skills/onboarding/runtime/critic.ts
11. 위험 & 미해결 질문
위험 (mitigation 동반)
- #2 abstention false-negative → ξ=0.35 보수, in-conversation abstain 차단(turnIndex ≥ 2일 때 clarify 안 함), 24h shadow 우선.
- #3 router 학습 데이터 편향 (현 정책 재현) → epsilon=0.05 exploration noise, 라벨 보강.
- #5 PRM 초기 라벨 부족 →
EVAL_PRM_MIN_LABELS=500미만 시 explanation-only 모드(forward score 미노출). - #6 PII 누출 →
UserMemoryStore.update에 anonymize wrapper 강제, audit job 매일 1회. - #1 SKILL.md 행동 변화 → 30일 long-soak, regression 시
AGENT_SKILL_MD_MODE=false즉시 rollback. - #7 propose 속도 저하 → T1→T2 자동(G1+G4만 요구)으로 가속 trade-off.
- #8b critic loop 폭주 →
AGENT_CRITIC_MAX_ITERATIONS=2강제, deadlineMs 체크.
미해결 질문 (구현 시작 전 사용자 확인 필요)
- 현재 label-archive 누적 라벨 수? PRM v1 위해 ≥ 500. 부족하면 review queue 가속을 Phase α에 추가.
- 재방문 user rate (chatroom 별 30일 동일 user turn 분포)? 메모리 ROI 산정.
- 로컬 embedding infra 추가 가능? (router v3, example-store v2 위해 Qwen-embed 64d). 외부 사용 가능?
- 다음 skill 후보 명세 — discord-help · community-rules · identity-room 중 어떤 것이 먼저? SKILL.md L3 resources 설계 영향.
12. 작업 체크리스트 (codex가 PR 단위로 진행)
Phase α (Week 1-2)
- PR-A1: TASK-α-2-01 (Hybrid Early Abstention) + TASK-α-FLAG-01 (RuntimeFlagManifest)
- PR-A2: TASK-α-8a-01 (Structured Tool Error Recovery)
- PR-A3: TASK-α-4a-01 (Programmatic Tool Calling 시범)
- Phase α Exit 게이트 검증 (Section 5)
Phase β (Week 3-5)
- PR-B1: TASK-β-1-01 (SKILL.md L1/L2/L3) — skill 추가 결정에 따라 최우선
- PR-B2: TASK-β-5-01 (PRM + Attribution)
- PR-B3: TASK-β-3-01 (Router v1 룰)
- PR-B4: TASK-β-6-01 (User Memory + Adaptive Few-shot)
- PR-B5: TASK-β-3-02 (Router v2 학습) — PRM 라벨 충분 시
- Phase β Exit 게이트 검증
Phase γ (Week 6+)
- PR-C1: TASK-γ-7-01 (Trust & Governance)
- PR-C2: TASK-γ-4b-01 (Tool Search Tool) — skill/tool 증가 시점
- PR-C3: TASK-γ-8b-01 (Multi-step Self-Critic)
- Phase γ Exit 게이트 검증
13. 비고
- 사용자가 highlight한
src/llm/routing.ts:24-25(return [llamaFallbackModel]) — 현 코드는provider === "llama"일 때 단일 fallback. TASK-β-3-01에서selectModelsForProvider시그니처에routingDecision을 inject하면 llama provider도 router의requiredTier에 따라 모델 선택 가능 (현재는 단일 모델이라 effective change 0, multi-llama 환경 대비 hook). - Phase α만으로 measurable 효과: cost −10%·quality flat·latency p50 −10%.
- 모든 PR은
evaluation-lively-island.md의 평가 인프라(shadow eval, tournament, drift)를 활용해 검증 가능. - Plan 원본(우선순위 분석·상세 트레이드오프)은
~/.claude/plans/papers-goofy-plum.md에 보관. 본 문서는 codex actionable view.