curviate-network
Grow and manage a LinkedIn network with the Curviate CLI. Covers `connect` (send with a note, sent, received, accept, decline, cancel), `profile follow`/`unfollow`, `profile relations`, `profile followers` and `profile following`, the de-duplication signal for a connect loop, and the invitation propagation delay. Use when sending or withdrawing connection requests, triaging received invitations, following a member, or listing connections and followers.
- License 许可
- MIT
- 最近更新
- 2026-11-12
- Stars
- ★ 3.3w
- Forks
- 3.9k
01 前言 Front Matter
- name
- curviate-network
- license
- MIT
- metadata
- { "tags": [ "LinkedIn", "CLI", "Agents", "Sales", "Recruiting", "Outreach" ], "author": "Curviate", "version": "0.1.0", "repository": "https://github.com/Curviate/curviate-plugin" }
- description
- Grow and manage a LinkedIn network with the Curviate CLI. Covers `connect` (send with a note, sent, received, accept, decline, cancel), `profile follow`/`unfollow`, `profile relations`, `profile followers` and `profile following`, the de-duplication signal for a connect loop, and the invitation propagation delay. Use when sending or withdrawing connection requests, triaging received invitations, following a member, or listing connections and followers.
03 SKILL.md 正文
Curviate: connections and following
An invitation is the most consequential low-effort write on LinkedIn: it is visible, it is attributable, and a withdrawn one still leaves a trace. Preview first, de-duplicate before sending.
Command surface established against CLI 0.33.0.
Before any command
npm install -g @curviate/cli && curviate --version # needs Node 18 or newer
curviate login --api-key <key> # or export CURVIATE_API_KEY
curviate account list --json # the acc_id for --account
- Credentials resolve flag > environment > stored profile (
CURVIATE_API_KEY,CURVIATE_BASE_URL,CURVIATE_ACCOUNT). --profile <name>picks the stored credential set;--account <acc_id>picks which connected LinkedIn account sends the invitation. Name it explicitly on every write: an invitation from the wrong account reaches a real person as that person.--previewbefore every write. It renders the resolved request (recipient, note, acting account) without sending. On a read command it is refused with exit2.--jsonon anything you parse;--fields a,b,cto project;--verbosewhen a slim response looks suspiciously empty.- Put global flags at the end of the command.
- Branch on the exit code, never on prose. See the table at the end.
- These are not retrieval-mode commands:
--mode/--max-ageare refused here withunknown flag, exit2(seecurviate-profilefor the four commands that do accept them).
connect: invitations
| Command | What it does | Confidence |
|---|---|---|
curviate connect <id> --note "<text>" |
Send a connection request. <id> is a slug, profile URL or member id. --note is capped at 300 characters by LinkedIn; omit it for a generic request. |
proven |
curviate connect sent |
Pending sent invitations only: accepted and declined ones are never returned, so this is a lower bound on what you sent. Use id with connect cancel; user.id identifies the recipient (the sent variant carries no public slug). created_at is the platform's own timestamp. There is no total count; use --all and count client-side. |
proven |
curviate connect received |
Pending received invitations only. user.public_identifier, display_name, first_name, last_name identify the sender. Use id with accept or decline. |
proven |
curviate connect cancel <id> |
Withdraw a sent invitation. <id> is the id from connect sent. |
proven |
curviate connect accept <id> |
Accept a received invitation. <id> is the id from connect received. |
wired, never live-fired |
curviate connect decline <id> |
Decline a received invitation. | wired, never live-fired |
connect accept and connect decline are preview-proven (--preview renders the correct call and
body), but have never been fired against a live account. They are real, correctly wired commands.
Run your own smoke test on an invitation you are prepared to act on before an unattended flow
depends on either.
Traps
- De-duplicate before sending. A request against a pair already mid-flow or already connected
returns
CONNECTION_REQUEST_CONFLICT, exit8. That is a clean, routine signal and the only reliable de-duplication marker for a connect loop, but checkprofile relationsandconnect sent --allfirst rather than probing with real invitations. - Invitations take roughly 10 to 30 seconds to appear on the recipient's side, and a very recent
one can take minutes to appear in
connect sent(LinkedIn-side indexing). An absence right after sending is not proof the send failed. Do not resend on that basis. --notetakes a string, not stdin. Write the note to a file, confirm it is non-empty, and pass--note "$(cat note.txt)". An empty note has gone out as a blank request.- There is no idempotency key and no server-side de-duplication. A send that times out may
already have landed; re-read
connect sentbefore re-issuing.
Following and connections
| Command | What it does | Confidence |
|---|---|---|
curviate profile relations |
Your first-degree connections. | proven |
curviate profile followers <id|me> |
A member's followers. me is always safe; another member's id also works. |
proven |
curviate profile following <id|me> |
Who a member follows. Self-only in practice, see below. | proven |
curviate profile follow <id> |
Follow a member. Sends a connection request instead if their profile is private. Bodyless write; resolves a slug or URL to the member id for you. | proven |
curviate profile unfollow <id> |
Unfollow. Idempotent. | proven |
profile followingis self-only. Passme. Any other member id (including your own explicit member id) returnsLINKEDIN_OPERATION_NOT_SUPPORTED, exit8. That is a permanent platform limitation, never retryable.profile followersdoes not share it.- Never write
profile me relations. The command isprofile relations. Older builds silently discardedrelationsand answered with your own profile at exit0; current builds exit2. - Following is the low-consequence alternative to an invitation: it needs no acceptance, is reversible, and still surfaces the member's posts in your feed.
A connect loop that behaves
# 1. Confirm you are not already connected or already mid-flow.
curviate profile relations --all --json # first-degree
curviate connect sent --all --json # pending invitations
# 2. Preview the exact request, then send it.
curviate connect <slug> --note "$(cat note.txt)" --account <acc_id> --preview --json
curviate connect <slug> --note "$(cat note.txt)" --account <acc_id> --json
On exit 8 with CONNECTION_REQUEST_CONFLICT, record the pair as already handled and move on. On
exit 13, stop the loop; see below.
Full command surface
<!-- generated: command surface, CLI 0.33.0 -->
Read from the CLI's own --help at version 0.33.0. Descriptions, traps and confidence
tags elsewhere in this skill are hand-written and carry the version they were established against.
Every command below that takes flags at all also accepts --account, --api-key, --base-url, --beta, --json, --preview, --profile, --timeout, --verbose.
| Command | Arguments | Flags |
|---|---|---|
curviate profile relations |
(none) | --fields, --limit, --cursor, --all, --max-pages, --page-delay |
curviate profile follow |
ID |
--fields |
curviate profile unfollow |
ID |
--fields |
curviate profile followers |
ID |
--fields, --limit, --cursor, --all, --max-pages, --page-delay |
curviate profile following |
ID |
--fields, --limit, --cursor, --all, --max-pages, --page-delay |
curviate connect |
ID |
--note |
curviate connect sent |
(none) | --fields, --limit, --cursor, --all, --max-pages, --page-delay |
curviate connect received |
(none) | --fields, --limit, --cursor, --all, --max-pages, --page-delay |
curviate connect accept |
ID |
(none) |
curviate connect decline |
ID |
(none) |
curviate connect cancel |
ID |
(none) |
<!-- /generated -->
Exit codes to branch on here
| Code | Meaning | What to do |
|---|---|---|
1 |
INTERNAL from the server itself: a genuine bug on the platform side. |
Worth one retry; if it repeats it is a bug to report, not a state to work around. |
2 |
Usage or invalid input, often raised before any network call. | Fix the invocation. Never retry unchanged. |
4 |
Not found: a wrong member or invitation identifier. | Re-resolve the id. |
5 |
Three causes, one code: read error.code. NO_ACTIVE_SEAT: the account is on no active seat. LINKEDIN_FEATURE_NOT_SUBSCRIBED: LinkedIn itself lacks the feature. BETA_NOT_ENABLED: the operation is beta-gated and this workspace has not opted in. |
Branch on error.code: the three fixes have nothing in common, and none is fixed by retrying unchanged. |
6 |
PLATFORM_RATE_LIMIT and its siblings. Carries retry_after in whole seconds. A response naming budgetRow means only that row is paused; every other row on the account keeps working. |
Back off and retry after that many seconds. On a named budgetRow, switch to other work on the account rather than backing off across the board. |
7 |
Transient platform fault: a request that got no response at all (network error, DNS failure, timeout) or one that came back as something other than a valid API answer. Carries retryLikelyToSucceed: true. |
Retry with backoff. |
8 |
Account or connection state. Read error.code: CONNECTION_REQUEST_CONFLICT is routine de-duplication; LINKEDIN_OPERATION_NOT_SUPPORTED is permanent; ACCOUNT_RESTRICTED, LINKEDIN_AUTH_FAILED and LINKEDIN_COOKIE_INVALID need a reconnect. |
Depends entirely on error.code; never treat the whole bucket as "reconnect the account". |
13 |
BUDGET_EXHAUSTED: a safety rule of your own refused the invitation, not LinkedIn. Read error.safetyReason: ceiling means the row named in error.budgetRow hit its configured limit; activity_window means the account is outside the hours it works in (no budgetRow on that one). Nothing reached LinkedIn and nothing was spent; no invitation went out. reset_at can be weeks out, and is null for allowances no clock frees (pending invitations, for instance, are freed by acceptances and withdrawals rather than by time). |
Do not back off and retry. error.safetyHint.parameter names the exact setting to change. Read quotas[] via curviate account get <acc_id> --json, then wait for the named reset, withdraw stale invitations, or change that setting. A retry loop only burns time. |
04 同仓库更多 Skills
wealthville-pool-scoring
Score DeFi liquidity pools before providing liquidity. Returns a 0-100 pool score and an Enter/Hold/Exit/Reduce/Avoid verdict with calibrated confidence for ~68,800 Solana pools and 575 EVM pools, plus a public miss-inclusive track record. Use when asked whether a pool is worth LPing into, which pools look best right now, or how accurate the scores have actually been.
technical-debt-auditor
Scans a codebase and produces a quantified technical debt report with a prioritized remediation plan. Analyzes dead code, TODO/FIXME/HACK markers, dependency staleness, complexity hotspots, code duplication signals, test coverage gaps, unused exports, and outdated patterns. Use this skill when the user asks to audit their codebase for tech debt, wants a health check of their code quality, says things like "where should I focus refactoring", "find code smells", "audit my codebase", "what technical debt do we have", "prioritize cleanup work", or asks about code quality metrics. Also trigger when the user mentions "code rot", "legacy cleanup", "refactoring priorities", or wants to understand the overall health of a project before a major effort. Trigger even if the user doesn't say "technical debt" explicitly — any request to systematically assess code quality and identify improvement priorities belongs here.
readonly-sql-guard
Make an agent's database access provably read-only. Use when an agent runs caller-supplied SQL against a production/ERP/analytics database (a "query" MCP tool, text-to-SQL, an admin console), or when reviewing DB-access code for write/DDL/side-effect/transaction-escape risks. Checks what a statement CALLS, not just its shape.
web-artifacts-builder
Suite of tools for creating elaborate, multi-component claude.ai HTML artifacts using modern frontend web technologies (React, Tailwind CSS, shadcn/ui). Use for complex artifacts requiring state management, routing, or shadcn/ui components - not for simple single-file HTML/JSX artifacts.
webapp-testing
Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs.
discernment-nudge
After you give a substantive answer or draft that the user may act on — advice or recommendations, drafted artifacts such as goals, plans, pitches, proposals, or emails, estimates or projections, analysis or interpretation of data, factual claims they may rely on, or a multi-step argument — invoke this skill BEFORE finalizing your reply and then, if it applies, append 2-3 short follow-up questions, each tied to something specific in what you just produced, that help the user check key facts, probe the reasoning or assumptions, and notice missing context. Do this at most once per conversation. Skip it when the user asked a trivial how-to or simple lookup, wants a purely educational explanation, asked you only to format, convert, or assemble a file from content they provided, is writing code they will run, is doing creative writing or casual chat, or already asked you to double-check, cite, or review — the skill file explains these boundaries and the exact output format.