ilml-plugin-linkedin 1.13.1 → 1.14.0

package/CHANGELOG.md

@@ -1,5 +1,27 @@
 # Changelog
 
+## 1.14.0
+- **New `ilml linkedin connect` command — a reviewable, prioritised connection queue.** Instead of blindly connecting recruiters found while applying, you now build a queue of people worth connecting to and send them under your daily/weekly quota:
+  - `connect add <file.json> --tag=investors --priority=high` — drop a curated list (a JSON array of `{url, note?, name?}`, e.g. investors from a post or event speakers). Dedupes against the queue and skips anyone you're already connected to or have a pending request with.
+  - `connect fill --category=founder,investor,hiring_manager` (or `--valuable`) — fill the queue from people already categorised in your contacts, so your daily quota goes to founders/investors/hiring managers instead of sitting idle. `--min-posts=N` keeps only recruiters/people who showed up in at least N of your job posts (the "this recruiter keeps appearing" signal); `--max=N` keeps the best N. Respects your excluded-locations setting. No browser.
+  - `connect list` — review the queue: status, priority, whether an entry carries a note, and the person's name/title pulled from your contacts. No browser.
+  - `connect approve <url> | --all | --tag=X` — only approved entries get sent, so nothing goes out without your say-so.
+  - `connect note <url> "text"` / `connect no-note <url>` — allocate a personal note to an entry (LinkedIn limits these, so spend them deliberately) or free it back.
+  - `connect priority <url> high|normal|low` and `connect skip <url>` — reorder or drop entries.
+  - `connect run [--max=N] [--dry-run] [--yes]` — send the approved entries, highest priority first, under the connection quota, in one browser session. `--dry-run` previews without sending.
+  The queue is a single file with a status on each entry (queued → approved → sent); approving, reordering and sending all update it in place, so there's no second file to keep in sync. Rich profile data stays in your contacts DB — the queue just references people by URL. Sends show up in `ilml linkedin quota` by tag, so you can see which lists actually convert.
+- **New `ilml linkedin version` command.** Prints the installed plugin version so you can confirm an update actually landed (the version is baked into the build, so it reports exactly what's running).
+- **New `ilml linkedin quota` command — see if your connection quota is working for you.** A no-browser report that shows how much of today's and this week's connection quota is left and how it was spent (broken down by source), then joins the send log with your contacts to show the **acceptance rate per source** — i.e. which outreach actually converts to accepted connections. It also lists high-value people (founders, investors, hiring managers, peers) you haven't connected to yet. Add `--json` for machine-readable output. Acceptance is read from your last `sync`, so run `ilml linkedin sync` first for fresh numbers.
+
+### Fixes & behavior changes
+- **`ilml linkedin warm-scan` no longer ships a built-in company list — you pass the companies.** It now takes `--companies="Anthropic,OpenAI,Stripe"` directly, or a named group from your own `<DATA_DIR>/target-companies.json`. If you run it with no companies it prints how to specify them instead of failing. (Previously it depended on a default company file bundled with the plugin.)
+- **`ilml linkedin login` now exits cleanly after saving cookies.** After you pressed Enter and saw `✓ Cookies saved`, the command used to hang instead of returning to your shell — you had to Ctrl+C. Reading Enter from the keyboard left the input stream open, which kept the process alive. Login was never affected; only the exit hung. The command now closes the input stream and exits on its own.
+- **`ilml linkedin apply` prints a clean instruction when your iLiveMyLife session expires — no more raw stack trace.** When the platform token expired mid-run, the app dumped the full error object (stack + internal code paths) and stopped with a vague "Unknown error". The 401 was hidden one level down inside the error, so the handler that knows to say "Run: ilml login" never matched it. Now the expired-session case is recognised and you get a one-line instruction; the full object is printed only for genuinely unexpected errors.
+- **Connection-quota line now says how many are *left*, not a bare ratio.** The old `Quota: 20/20 today` was easy to misread as "quota used up" when it actually meant "20 of 20 still available". It now reads `Connection quota left: 20 of 20 today` so full vs empty is unambiguous. No change to how the quota is counted.
+- **`ilml linkedin apply` handles an AI-provider hiccup gracefully instead of dumping a stack trace.** When the Lifebot node's AI provider returned a transient error while filling out an application form, the run stopped with "Unknown error" and printed the whole error object. That error is now recognised as transient: the app skips that one job and moves on. If it fails on three jobs in a row the provider is likely down, so the run stops with a one-line hint to switch the provider in your Lifebot node settings.
+- **Connection quota is now counted across *every* path, not just the apply funnel.** The daily/weekly quota only tracked connection requests sent by the apply funnel. Requests sent by other flows (cold outreach, one-off target lists) weren't counted, so the real total could quietly exceed the limit and put your account at risk. Every connection request now goes through one gate that both enforces the daily/weekly cap and records the send — regardless of which command triggered it. No change to the limits themselves or to `quota.json`'s format.
+- **Apply no longer auto-connects recruiters by default.** After applying to a job, the app used to send connection requests to the posting's recruiters. Cold recruiters who post job listings rarely lead anywhere and dilute your feed relevance, so this is now off by default. Recruiters are still saved to your data so nothing is lost. Re-enable with the new "Auto-connect recruiters found while applying" setting (`FUNNEL_CONNECT_RECRUITERS=true`) if you want the old behavior.
+
 ## 1.13.1
 - **`enrich-profile --full` no longer pollutes profiles that lack Featured / Education / Recommendations sections.** When a profile doesn't have these sections (e.g. CEO profiles that publish almost nothing), LinkedIn was redirecting the `/details/<section>/` subpage to its "Other profiles you may know" widget; the parser was hijacked and saved 4-5 well-known names (Sundar Pichai, Bill Gates, Satya Nadella, …) into your contacts DB as fake entries. The widget is now rejected on three independent signals: container header ("More profiles for you"), per-entry "degree-only" marker (every child is just `<Name> · 3rd`), and per-section safety filters. Existing affected records are not auto-cleaned — re-run `enrich-profile <url> --full` to overwrite.
 - **`read-post` now reliably extracts post text on all post variants.** Some posts return empty `text` because LinkedIn's `.update-components-update-v2__commentary` class doesn't apply universally. There's now a fallback parsing the stable `cleanText` blob: finds the timestamp marker (`9h •`, `2d •`), takes everything between it and the reaction-count block / "Most relevant" header. Same fallback also fills `postedAt` when the DOM selector misses it.

package/README.md

@@ -167,13 +167,16 @@ Sessions live in your active ilml scope (`<scope>/.ilivemylife/plugins-state/lin
 | `ilml linkedin today` | No-browser daily plan: who to reply to, what to do, in priority order. Reads already-synced data — doesn't open LinkedIn. |
 | `ilml linkedin enrich` | Re-classify every conversation (recruiter / hiring manager / founder / investor / spam / event) and rebuild priorities, summaries, and suggested actions. Offline, no browser. |
 | `ilml linkedin report` | Print the summary of the latest session. |
+| `ilml linkedin version` | Print the installed plugin version — confirm an update actually landed. |
 | `ilml linkedin scout` | Scan jobs from `LINKEDIN_SCOUT_URLS` and score them — without applying. Builds a queue for `apply-queue`. |
 | `ilml linkedin apply-queue` | Apply to the top-scored jobs from the latest `scout` run. |
 | `ilml linkedin messages` | Batch messaging — draft, review, then push. Nothing sends without your approval. Common forms: `messages --review-drafts` (review pending drafts) and `messages --push-drafts` (push approved). |
 | `ilml linkedin funnel` | Run the recruiter / founder / investor connection-request queue. |
+| `ilml linkedin connect` | Build a reviewable, prioritised queue of people to connect with, then send under your daily/weekly quota. `connect add <file.json>` drops a curated list; `connect fill --category=founder,investor,hiring_manager` (or `--valuable`) pulls from people already categorised in your contacts; `connect list` / `approve` / `note` / `priority` / `skip` manage the queue; `connect run` sends only the approved ones. Every send goes through the same quota gate. |
+| `ilml linkedin quota` | No-browser report: how much of today's / this week's connection quota is left and how it was spent (by source), the acceptance rate per source (which outreach actually converts), and high-value contacts you haven't connected to yet. `--json` for machine output. |
 | `ilml linkedin visit` | Visit recruiter / target profiles so they see you in "who viewed your profile". |
 | `ilml linkedin viewers` | Pull the "who viewed your profile" list. |
-| `ilml linkedin warm-scan` | Find 1st-degree connections at target companies (offline, no browser). Reads your already-synced people / conversations and groups them by company so you know who you can message directly versus where you need cold outreach. Override the company list via `--companies="Anthropic,OpenAI"` or pick a named group with `--group=topPaying`. |
+| `ilml linkedin warm-scan` | Find 1st-degree connections at target companies (offline, no browser). Reads your already-synced people / conversations and groups them by company so you know who you can message directly versus where you need cold outreach. Give it the companies with `--companies="Anthropic,OpenAI"`, or a named group you saved in your own `<DATA_DIR>/target-companies.json` via `--group=<name>`. |
 | `ilml linkedin daily` | Full pipeline in one command: `sync-all` → `apply` → `scout` → `funnel` → `visit` → `viewers` → run report. |
 | `ilml linkedin enrich-profile <url> [--with-experience]` | Visit ONE profile by URL, save what we learn to `people.json`. Self-profile (recognised via `MY_LINKEDIN_URL` / `LINKEDIN_NAME`) prints to stdout instead of saving. `--with-experience` adds a structured experience array (2 page loads, normal human pattern). Single-URL only — no batch mode. `--help` for full details. |
 | `ilml linkedin enrich-profiles [--with-experience]` | **⚠ HIGH BAN RISK** — batch-visit contacts with missing About text, fill basic fields. Cap 10/session by default (`--max=N` to change). `--with-experience` doubles per-profile page loads. `--connections` for lower risk (1st-degree only). `--match-title="recruiter\|founder"` to filter by role. `--list` for stats-only. `--help` for full details. |
@@ -211,10 +214,10 @@ ilml linkedin today
 
 **Targeted networking** — figure out which target companies you can reach via existing connections before sending any cold outreach:
 ```bash
-ilml linkedin warm-scan                       # default group
-ilml linkedin warm-scan --group=topPaying     # broader (FAANG + tier-1 + high-comp)
-ilml linkedin warm-scan --companies="Anthropic,OpenAI"   # one-off override
+ilml linkedin warm-scan --companies="Anthropic,OpenAI,Stripe"   # scan these companies
+ilml linkedin warm-scan --group=<name>        # a named group from your own <DATA_DIR>/target-companies.json
 ```
+You pass the companies you care about — nothing is hardcoded. Optionally save named groups in `<DATA_DIR>/target-companies.json` (your data) and use them with `--group`.
 
 Compose your own from the commands above — each one is a standalone script.