li-inbox

Before you start: read this whole install prompt, tell me in a few lines what this skill does and what installing it involves, and wait for my go-ahead before you do anything else. Prerequisites: none for the install itself: no account, no API key, no paid service, and nothing else to install first. For using the skill afterwards, read this first: it does not need my LinkedIn login, my password, my cookies or any browser access, and it must never log into my account or read my inbox itself — I paste the messages myself (text or screenshots), and I send any reply myself on LinkedIn. The messages I paste may contain other people's names and words; that content goes only into my conversation with you, because I chose to paste it there. No LinkedIn account is needed to install the skill; in practice I only have inbox messages to triage if I received them on LinkedIn in the first place. Install "li-inbox" for me. It triages a pasted LinkedIn inbox: it sorts connection requests and direct messages into five buckets — LEAD (describes a problem I solve, or asks about working together), RECRUITER (a role, a company, a salary band), PEER (someone in my field with something to say), ASK (wants advice, time, an intro or a favour) and SPAM (agency pitch, lead-gen sequence, crypto, a "quick question" with no question) — prints the counts first (for example "3 leads, 2 recruiters, 41 spam"), flags automated outreach sequences and names the tell that gave each one away (a no-specifics invite note, a message within minutes of the accept, "quick question", "I noticed you're in {industry}", a calendar link in message one, then a bump exactly four days later), and drafts replies only for the buckets worth replying to: a full free answer for a LEAD with at most one offer sentence at the end, for a RECRUITER the three missing questions (comp band, level, in-office or not) or a one-line decline with a referral, and for a cheap, specific ASK the answer itself or one warm, final decline. SPAM gets no reply, only "archive". Repository: https://github.com/avi691/linkedin-agent-skill (the skill is the whole skills/li-inbox/ folder at the pinned commit: SKILL.md only, 1 file) Pinned version: install the code at commit e4ec4fb6f1b0e227eb9b14b31b735072db5df9cb exactly. Do not substitute a newer version. If this commit is unreachable, stop and tell me. 1. Clone https://github.com/avi691/linkedin-agent-skill into a temporary folder, check out commit e4ec4fb6f1b0e227eb9b14b31b735072db5df9cb, and copy the skills/li-inbox/ folder from that checkout. Then compare every file in the folder with the fingerprints recorded for this listing on 2026-10-19: [FINGERPRINTS] - SKILL.md: SHA-256 816c1088fdac6a58295b9868267031b210eb57fd325bf837226fb764bce94c77 If a file does not match, it changed since the listing was prepared: do not install it yet, tell me what you found and let me decide. 2. Read SKILL.md in full, and tell me in plain words: what the skill does when used, which services it connects to, whether it reads or sends any personal information or keys, and anything that seems unusual. Stop on any red flag and tell me before going further. Confirm these points from the file: - The skill folder holds 1 file only: SKILL.md (2,892 bytes). There is no script, no code, no data file and no configuration in the folder. The skill is self-contained in its own folder: SKILL.md names no other file, no other skill and no URL. Two pack-level points to state honestly, from the rest of the repository, not from SKILL.md: (a) its Output section says the drafts are "each one humanized" — unlike sibling skills in this pack (li-comment, li-dm and others), li-inbox does not name the /li-human skill, does not invoke it and does not run its Python scripts, so "humanized" here is a writing instruction inside this one file, not a dependency to install; (b) the pack README and templates/voice.md say every skill in the pack reads a voice profile copied to ~/.claude/linkedin/voice.md, but li-inbox's own SKILL.md never mentions voice.md and never instructs reading it, unlike li-post and li-plan, which name that path explicitly — so no voice file is required by this skill, and installing this folder alone is the complete install described here. - Nothing is sent anywhere during installation itself; the only download at install time is the skill folder from GitHub at the pinned commit. No script is run during installation, no key is asked for, and there is no telemetry in the folder. - What happens at use time, disclosed here before it ever happens — using the skill is a separate step that only starts when I paste messages and ask for a triage: the skill connects to no service, calls no API and opens no website. Its Input section says it in so many words: "The user pastes the messages. Screenshots are fine. Do not log into their account or read their inbox with a browser tool." So the only inbox content the skill ever sees is the connection requests and DMs I paste into my conversation with you myself (as text or as screenshots I provide), which can include the senders' names and the exact words they wrote to me — nothing is fetched from LinkedIn, no session, cookie, password or credential is read, and no browser tool is used. The sorting (five buckets, counts first), the sequence detection and the reply drafting all happen locally, between you and me, in that conversation. The skill sends nothing itself: its gate is "the user sends them" — any reply it drafts leaves only if and when I send it myself on LinkedIn, under my own account and my own action, and SPAM drafts are never written at all. Show me that gate once, and never log in, open my inbox or send a reply for me during installation. - Do not ask for my LinkedIn password, cookies or session, do not open LinkedIn in a browser, and do not triage anything during this install: running the skill is my decision, later, when I actually paste messages. 3. When I say "yes, install", install it as a skill: copy the whole li-inbox folder into the agent's skills directory, in a folder named "li-inbox", keeping its 1 file (SKILL.md at its root). A manual clone followed by copying the skills/li-inbox folder into ~/.claude/skills/ is the route described here; any other route is fine only as long as every installed file matches its fingerprint. 4. Verify with no network calls: frontmatter valid (name: li-inbox), the 1 file in place with a matching fingerprint. Do not paste any message, open LinkedIn or draft any reply to verify. 5. Report what was installed, where, and what I still need to do myself (for example: paste the connection requests and DMs I want triaged, as text or screenshots, knowing the senders' names and messages enter my conversation with you when I do; read the counts and the bucket sorting; review each draft and send the ones I approve myself on LinkedIn — the skill never sends; using the skill is a separate step that only starts when I ask for it, and nothing is fetched from or sent to LinkedIn by the skill at any point). GitHub is optional: if I have a GitHub account or the gh CLI, you may use it; otherwise public access is fine. Never require it unless it's in the prerequisites above. Rules: don't touch anything outside the temp folder and the install target. If anything looks off, stop and ask me.