← ClaudeAtlas

job-digestlisted

Use this skill whenever the user asks about job/internship postings from subscribed sources — «что нового по вакансиям», «есть что в Changellenge», «покажи стажировки за неделю», «добавь канал с вакансиями», «подпишись на ещё один tg-канал», or wants the daily vacancy digest set up, changed, or debugged. It pulls Telegram channels and RSS feeds through a local RSSHub, filters them by keywords, and drops a daily digest into the user's Mail.ru folder «Вакансии».
Sapar-hub/sapar-skills · ★ 0 · AI & Automation · score 72
Install: claude install-skill Sapar-hub/sapar-skills
# job-digest Pipeline: `feeds.txt` → RSSHub (`127.0.0.1:1200`, docker) → `scripts/digest.py` (keyword filter, dedup in sqlite) → IMAP APPEND into the Mail.ru folder `Вакансии` → systemd `--user` timer once a day. Files: - config: `~/.config/job-digest/feeds.txt`, `keywords.txt` (regexes, one per line) - state: `~/.local/state/job-digest/items.db` (every item seen, `sent_at` once mailed) - credentials: `MAILRU_USER`, `MAILRU_APPPASS` (Mail.ru app password) from the environment (fish universal vars on the author's machine) or `~/.config/job-digest/env`. Never print, echo, or commit the password. ## Answering "what's new" Read the local base first; it needs no network: ```bash python3 -I scripts/digest.py show --days 7 # what passed the filter python3 -I scripts/digest.py show --days 7 --all # including noise ``` Summarise per vacancy (role, company, conditions, link) and point out the ones that match the user's profile: DS / ML / data analytics / product analytics internships and junior roles, Moscow or remote. For fresh data, run `digest.py run` first: it fetches and mails new items. ## Adding a source - Telegram channel: append `http://127.0.0.1:1200/telegram/channel/<username>` to `feeds.txt` (the channel must be public: `https://t.me/s/<username>` opens without login). - Any other site: find the RSSHub route at https://docs.rsshub.app or use the site's own RSS/Atom URL. - Check with `digest.py feeds`, then `digest.py run --dry-run`. - The first run o