원클릭으로 Manus에서 모든 스킬 실행

poly-copy-trading

스타4

포크5

업데이트2026년 5월 10일 21:04

Cogni poly copy-trade mirror pipeline specialist. Load when working on the mirror loop itself — `mirror-coordinator`, `wallet-watch`, `poly_copy_trade_{targets,config,fills,decisions}` tables, poll cadence, v0 live-money caps, RLS on copy-trade tables, shared poller refactor (task.0332), status-sync / sync-truth cache (task.0328), v1 hardening bucket (task.0323), or Phase 4 streaming prep (task.0322). Also triggers for: 'mirror this wallet', 'why didn't the mirror fire', 'flip copy_trade_config', 'tracked wallet add', 'mirror skip reason=already_placed noise', 'fills ledger status drift', 'cap exceeded daily', 'WebSocket streaming poly', 'target ranker'. For provisioning trading wallets / CLOB creds / CTF approvals see `poly-auth-wallets`; for CLOB order semantics / Data-API / wallet screening see `poly-market-data`.

설치

Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.

Manus에서 실행

출처

cogni-dao

cogni-dao/cogni

GitHub 저장소 열기 Creator 저장소 보기

다운로드

Manus에서 실행

관련 직업SOC

SOC 직업 분류 기준

데이터베이스 아키텍트컴퓨터 및 수학직·SOC 15-1243

SKILL.md

readonly

이 저장소의 다른 Skills

같은 저장소

manage-node-envs

cogni-dao/cogni

Manage which deploy-environments a Cogni node lives in (candidate-a / preview / production) via the operator app — the full request→owner-approve→verb→merge→verify loop for env (deploy-topology) membership. Use whenever an agent needs to add a node to an env or remove it from one, is asked to 'give this node a preview deploy', 'promote this node into production topology', 'take a node out of preview', 'stop deploying X to candidate-a', 'change a node's deploy reach / env set', 'onboard my node to an environment', or hits a 403 authz_denied / 503 authz_unavailable on POST /api/v1/nodes/{id}/envs. Also triggers on 'env_manager', 'can_manage_envs', 'env-membership PR', 'node deploy topology', 'catalog envs: line'. This is distinct from FLIGHTING a build (candidate-flight / can_flight) and from PROMOTING a sha (promote / can_promote_production) — env-membership changes WHERE a node is deployed, not WHICH build runs there. Do NOT use for provisioning a whole environment (that is provision-env), for secret-writes (

2026-07-174

node-to-candidate-a

cogni-dao/cogni

Take a Cogni node from NOT-deployed to live-and-validated on candidate-a, end to end, as a node-contributor agent — RBAC request → owner approve → env-membership verb (add candidate-a) → merge → Argo auto-reconcile → agent-api-validation. This is the cold-start runbook: a fresh agent with only its COGNI API key + a node id can follow it start to finish. Use when asked to 'get my node onto candidate-a', 'deploy node X to candidate-a', 'prove a node deploys e2e', 'add a node to the candidate environment', or to validate the env-membership verb against a real deployment. Distinct from node-setup (which BIRTHS a node into the registry — a prerequisite here) and from flighting/promoting a BUILD (candidate-flight / promote — which pick which image runs in an env you're already in). This picks the ENV itself, then proves the node actually deployed and answers.

2026-07-094

test-expert

cogni-dao/cogni

Authoritative reference for the cogni-template test pyramid and production-like e2e test topology — enforcement vs test layers, vitest configs, infra prereqs, GitHub/DoltHub test org boundaries, the cogni-operator-test app, what agents can/can't run locally, coverage tracking, and the non-obvious gotchas that bite every time. Use this skill whenever the user is writing a new test, asking "which layer does this belong in", debugging a flaky test, hitting CWD/env-loading issues, deciding between unit/component/stack/external, running any `pnpm test:*` or `pnpm arch:check` / `pnpm lint` command, working with testcontainers, touching fake adapters or `APP_ENV=test`, troubleshooting skip-gates, seeing errors from `.env.test` not loading, wondering about coverage or whether a test will run in CI, or trying to decide whether to run stack/e2e locally vs defer to CI. Also trigger when the user mentions mocking the database, Privy/GitHub App construction errors in tests, the smee proxy, `pnpm test:smee`, validating the

2026-07-094

devops-expert

cogni-dao/cogni

CI/CD architecture reviewer and DevOps advisor for monorepo pipelines, branch workflows, and deployment strategy. Use this skill when reviewing GitHub Actions workflows, designing promotion pipelines, evaluating branch protection rules, reviewing deploy scripts, auditing CI/CD for anti-patterns, or making decisions about build/test/deploy strategy. Also triggers for: 'review this workflow', 'is this CI pattern correct', 'branch strategy', 'deploy pipeline design', 'release promotion', 'should we use PRs for this', 'CI is slow', 'pipeline gap analysis', 'provision', 'deploy script'.

2026-06-304

payments-expert

cogni-dao/cogni

Use whenever working on Cogni USDC credit top-ups, payment intent creation, payment activation, operator wallet funding, 0xSplits allocations, x402 outbound provider payments, Hyperbolic, EIP-3009 signatures, retired OpenRouter/Coinbase top-ups, system-tenant revenue share, repo-spec payment config, live payment rail guards, payment CI failures, or debugging money-loop behavior. Trigger on phrases like payments, payment rail, credits top-up, USDC, Split, 0xSplits, splitHash, x402, Hyperbolic, facilitator, signX402Payment, EIP-3009, OpenRouter funding, Coinbase Commerce, operator wallet, Privy wallet, markup_factor, revenue_share, system bonus, provider top-up, settlement_skipped, SPLIT_CONFIG_MISMATCH, activate payments, node payments, or "is this payment behavior validated". This skill is the launching point for future payments work and debugging.

2026-06-284

conductor-worktree-setup

cogni-dao/cogni

Use when setting up, reviewing, or changing Cogni Conductor worktree bootstrap, including conductor.json, scripts/conductor-worktree-setup.sh, secret/auth symlinks, and AI launch-pack references for human or AI developers.

2026-06-274

name

poly-copy-trading

description

Poly Copy-Trading Pipeline

You are the expert for the mirror loop — everything that turns "target wallet traded" into "our wallet placed a mirror order." Trading-wallet provisioning, signing, and CLOB semantics live in sibling skills.

Architecture in one pass

wallet-watch source                   (WS wake-up + Data-API drain, or legacy source in old branches)
    │
    ▼
features/wallet-watch                 (normalize → Fill; no decisions, no writes)
    │
    ▼
mirror-pipeline.ts::runMirrorTick     (dedup, target policy, position context)
    │
    ▼ INSERT_BEFORE_PLACE              (correctness gate — at-most-once)
poly_copy_trade_fills row lands       (status=pending, client_order_id set)
    │
    ▼
PolyTradeExecutor.placeIntent         (authorizeIntent → CLOB adapter)
    │
    ▼
ledger updated + decision row written (placed | skipped | error)

Invariant order matters. Flipping INSERT and PLACE means a successful CLOB submit whose ledger row never committed → next poll double-mirrors. Never reorder.

Durable flowchart: Poly Order And Position Lifecycle § Mirror Decision To Limit Order.

Key code landmarks

nodes/poly/app/src/features/copy-trade/mirror-pipeline.ts — runMirrorTick glue
nodes/poly/app/src/features/copy-trade/plan-mirror.ts — pure mirror decision policy
nodes/poly/app/src/features/wallet-watch/ — wallet observation source, no decisions
nodes/poly/app/src/bootstrap/jobs/copy-trade-mirror.job.ts — poll shim + target-config hydration
nodes/poly/app/src/features/copy-trade/target-source.ts — dbTargetSource + envTargetSource; target policy fields
nodes/poly/packages/db-schema/src/copy-trade.ts — schema for poly_copy_trade_* tables
nodes/poly/app/src/features/trading/order-ledger.ts — ledger writes + sync-truth reads

Tables (RLS-scoped, per-tenant)

Table	Purpose
`poly_copy_trade_targets`	Per-tenant tracked target wallets plus target policy fields. RLS via `created_by_user_id`.
`poly_copy_trade_fills`	The ledger. Row per decision. `idempotency_key = keccak256(target_id + ':' + fill_id)`.
`poly_copy_trade_decisions`	Append-only audit log of every `decide()` outcome. Never updated or deleted.

dbTargetSource.listForActor(actorId) — RLS-clamped via appDb, used by per-user routes. dbTargetSource.listAllActive() — under serviceDb, the ONE sanctioned BYPASSRLS read across tenants. Only called by the mirror-poll enumerator in container.ts. If you find a second caller, that is a bug.

Runtime config — candidate-a

Enable switch: there is no per-tenant kill-switch table. An active target row plus active wallet connection plus active grant is the opt-in gate. Remove or disable the target row to stop mirroring that wallet.

Poll cadence: 30s. Warmup backlog: 60s. Hardcoded in copy-trade-mirror.job.ts. Bounds our latency floor at ~30-60s (task.0322 P4 swaps to CLOB WebSocket).

Target policy: poly_copy_trade_targets carries mirror_filter_percentile and mirror_max_usdc_per_trade. The enumerator threads these through buildMirrorTargetConfig into MirrorTargetConfig.sizing.

Grant caps: poly_wallet_grants carries per_order_usdc_cap, daily_usdc_cap, and hourly_fills_cap. These are enforced by authorizeIntent, downstream of planMirrorFromFill.

Tracked wallets — add via dashboard + or POST /api/v1/poly/copy-trade/targets; remove via − or DELETE /api/v1/poly/copy-trade/targets/[id]. Never seed wallets via env vars (the Phase-A envTargetSource exists for local-dev only).

Observability — mirror signals

Signal	Where	Good state
`poly.mirror.poll.singleton_claim`	Loki	Fires exactly once per pod start
`poly.wallet_watch.fetch`	Loki	`raw=N, fills=N, phase=ok` when a drain runs
`poly.wallet_watch.ws.wakeup_total`	Metrics	Increments when watched assets trade
`poly.mirror.decision outcome=placed`	Loki	Emitted when mirror fires
`poly.mirror.decision outcome=skipped reason=already_placed`	Loki	Dedup. Noisy — reducing noise is in task.0323 §1.
`poly.mirror.decision outcome=skipped reason=position_cap_reached`	Loki	Hitting target-policy cap. Expected under load; investigate if spiking.
`poly.mirror.poll.tick_error`	Loki	ZERO. Any hit = bug.
`poly_copy_trade_fills`	poly DB	Row per mirror decision

Status-sync gotcha (task.0323 §2 / task.0328): poly_copy_trade_fills.status=open is written at INSERT time and never re-read from CLOB. Actual CLOB state may be filled, canceled, or partial. Don't trust the ledger's status column alone — cross-check Data-API /positions?user=<addr> or the synced_at staleness window exposed via /api/v1/poly/sync-health.

MCP-down fallback: scripts/loki-query.sh '{env="candidate-a",service="app",pod=~"poly-node-app-.*"} | json | event=~"poly.mirror.*"' 30 — see top-level poly-dev-manager skill for details.

Idempotency + fill_id — frozen

fill_id = data-api:<tx>:<asset>:<side>:<ts> — assembled in polymarket-source.ts. Shape is frozen. Phase 4 adds clob-ws:<…> as a sibling scheme; never mix schemes within one fill.
idempotency_key = keccak256(target_id + ':' + fill_id) — written to ledger and submitted as client_order_id to CLOB. Invariant: two poll ticks seeing the same on-chain fill produce the same key → CLOB dedups → at-most-once.
The target_id in the key is the UUIDv5 derived from target_wallet, not the DB row PK (UUIDv4). This was the source of revision-1 bug #2 on task.0318 Phase A — HTTP routes had a drift between target_id (UUIDv5 for ledger correlation) and params.id (UUIDv4 DB row PK for DELETE). Same term, different space.

v1 + Phase 4 roadmap pointers

task.0323 v1 hardening — cursor persistence, CTF SELL (wallet-owned, overlaps with poly-auth-wallets), status-sync, metrics, alerting. In Review at time of writing.
task.0328 sync-truth cache — DB-as-CLOB-cache with synced_at + /sync-health. Done.
task.0332 shared poller — replace per-wallet setInterval with one poll loop + TargetSubscriptionRouter. Blocks Phase 3 multi-tenant scale. Needs Design.
task.0322 Phase 4 — CLOB WebSocket dual-path ingestion + hot signer + target ranker + counterfactual PnL. Needs Design.

Active bugs that affect this pipeline

bug.0335 — shared operator BUY empty reject on candidate-a. Every autonomous mirror attempt rejected with empty CLOB response. Likely operator-wallet balance / allowance / keys — not a code bug in this pipeline. Ground-truth in poly-auth-wallets. Until resolved, candidate-a mirror attempts always fail at the CLOB submit step.
bug.0329 — SELL on neg_risk market empty reject. Any position opened on neg_risk becomes roach-motel until resolution. Blocks close-position. Root cause: missing CTF approval (see poly-auth-wallets).

Anti-patterns specific to the mirror

"I placed a trade from my own per-tenant wallet and the mirror should see it." No — the mirror watches the TARGET wallet, not your trading wallet. Placing from yourself validates CLOB + signing, not the mirror.
Adding a second BYPASSRLS read somewhere other than the container.ts enumerator. Cross-tenant reads are a cross-tenant isolation hole; the enumerator is the audited seam.
Generalizing scripts/experiments/place-polymarket-order.ts (scope-narrow $1 post-only dress-rehearsal) instead of using privy-polymarket-order.ts. The dress-rehearsal has deliberate guardrails.
Smuggling P4 scope (WebSocket, ranking, streaming) into v0 or v1. Keep the fill_id shape + idempotency formula fixed.
Silencing outcome=skipped reason=already_placed at the coordinator layer. It's noise at the Loki level but correctness-preserving at the code level. Fix per task.0323 §1 (batch / sample), not by removing the skip branch.
Changing v0 caps via kubectl set env. Argo reverts on next sync. Caps are in code. task.0347 is the work item to lift them to tenant-config.

Enforcement rules

INSERT_BEFORE_PLACE is the correctness gate — never reorder.
fill_id and idempotency formulas are frozen.
Per-tenant RLS on poly_copy_trade_* tables must stay intact. Any new table joins through billing_accounts.owner_user_id EXISTS-pattern (same shape as llm_charge_details), not direct created_by_user_id coupling.
Caps are defense-in-depth. Even when task.0347 lifts them to per-tenant config, the code must enforce whichever is smaller: configured cap or hardcoded safety cap.