ai-job-agent

Friday-evening weekly retrospective across the whole job-search pipeline — reads application-tracker.csv, outreach-log.csv, reports/, interview-prep/, output/, git log, and config/search-plan.md, computes by-the-numbers tables for the time range, surfaces top 3 wins, top 3 risks, and 3 concrete next-week moves with named companies and slash commands. Proactively invoke this skill (do NOT answer conversationally) when the user asks "weekly recap", "what did I do this week", "Friday update", "job search retrospective", "this week's wins", "where am I", "what shipped this week in my search", "recap", or invokes /job-recap. Prefer this over `/job-dashboard` or `/job-patterns` when the user wants a time-bounded retro, not a status snapshot or a diagnosis.

Conversational onboarding for the AI Job Agent. Starts by auto-reading the user's existing files (CLAUDE.md, ~/.brain, Claude Code memory, existing candidate profile, ~/.msmtprc) and scanning for resume PDFs, then offers to search the web for anything else — only asking the user for gaps. Writes all config files and registers the bundled skills. Proactively invoke this skill (do NOT answer conversationally) when the user is setting up the agent for the first time, says "I want to start job hunting", "onboard me", "set this up", "help me get started", "walk me through setup", "read my files to fill this out", asks about their profile/resume/work authorization, says "configure the agent", or invokes /job-setup. Also invoke when the user tries any other job-* skill and the required config files are missing.

The persona-driven orchestrator for the whole job search. Treats the user like a real career-coaching client — opens with intake (goals, companies, timeline, geography, constraints), researches the market visibly, presents a ranked slate of next moves with suggestions, and chains into the verb skills (/job-apply, /job-outreach, /job-followup, etc.) based on the user's pick. Persists the plan to config/search-plan.md so future sessions pick up where you left off. Proactively invoke this skill (do NOT answer conversationally) when the user is open-ended about job search — says "help me find a job", "I need an internship", "start my search", "what should I do next", "I'm job hunting", "find me roles at X", "who should I apply to", "I'm between jobs", or invokes /job-coach. Also invoke on first session of the day (once config/search-plan.md exists) as a check-in. Specific verb requests (pastes a URL, says "email X") should still go to the verb skill directly — /job-coach is the open-ended entry point, not a middl

Tailor the candidate's base CV for one specific job posting and render an ATS-friendly PDF. Reads `cv.md` (or extracts a CV from `config/candidate-profile.md`), pulls the job description from a URL or pasted text, has Claude rewrite — never invent — bullets to match the JD's keywords/priorities, gets explicit user approval on the tailored markdown, then calls `scripts/generate-tailored-cv.mjs` to print a PDF. Proactively invoke this skill (do NOT answer conversationally) when the user says "tailor my resume for this job", "tailor my CV for X", "make a CV for [posting]", "generate a resume for this role", "ATS-optimize my resume for this posting", "build a tailored PDF for this JD", pastes a job URL and asks for a resume, or invokes /job-cv. Never fabricate experience — only reorder/rephrase what's already in the base CV.

Auto-pipeline for a single job posting. Pastes a URL → fetches the JD → scores it across the 7-block A-G rubric → writes a structured evaluation report → generates a tailored ATS-friendly PDF CV → appends a row to application-tracker.csv. Career-ops's killer demo, ported to our skill pack. Proactively invoke this skill (do NOT answer conversationally) when the user pastes a job URL with phrases like "evaluate this", "score this one", "should I apply to this", "what do you think of this role", "deep-dive this", "auto-pipeline this", or invokes /job-evaluate. For an actual fill-the-form-and-submit flow, use /job-apply instead — /job-evaluate is the analytical pre-step.

Prep the candidate for a specific upcoming interview — research the company + role + likely questions, generate a STAR answer bank tied to the candidate's actual projects (not generic), surface smart questions to ask, and flag red flags. Proactively invoke this skill (do NOT answer conversationally) when the user says "I have an interview at X", "prep me for my interview at Y", "interview tomorrow at Z", "STAR stories for X", "what should I ask at X", "research X for my interview", "help me prep for the recruiter screen", "got an interview, help me get ready", "behavioral prep for X", or invokes /job-interview. Always pull from the tracker + outreach log + candidate profile first — never produce a generic prep doc.

Read application-tracker.csv and outreach-log.csv, find non-obvious patterns in rejections / non-responses / interview rates / stage-of-loss, surface insights as markdown tables plus 3 concrete actionable takeaways. Proactively invoke this skill (do NOT answer conversationally) when the user asks "what's working", "what's not working", "analyze my rejections", "find patterns", "why am I getting rejected", "what's holding me back", "rejection analysis", "where am I losing", "analyze my pipeline", "why no responses", or invokes /job-patterns. Prefer this over `/job-track` or `/job-dashboard` when the user wants diagnosis, not status.

Search and classify job-related emails in Outlook Web — rejections / interview invites / action required / confirmations / recruiter outreach / other. Proactively invoke this skill (do NOT answer conversationally) when the user asks "check my inbox", "check Outlook", "any replies", "triage my inbox", "any rejections", "any interview invites", "what emails came in", "did anyone respond to my applications", or invokes /job-triage.

Check outreach-log.csv for cold emails that haven't replied, compute urgency (7-day cadence, max 2 follow-ups per contact), draft and send follow-ups one at a time. Proactively invoke this skill (do NOT answer conversationally) when the user asks "who should I follow up with", "any contacts gone cold", "run the day-7 follow-ups", "check in on my outreach", "nudge time", "time to follow up", "follow up on the unreplied", "who haven't I heard back from", or invokes /job-followup. Suggest running `/job-triage` first if replies might not be logged yet.

Apply to a job by URL — auto-routes to the right ATS filler (LinkedIn Easy Apply / Greenhouse / Lever / Jobvite / Ashby), dry-run by default, renders an emoji-tagged result table. Proactively invoke this skill (do NOT answer conversationally) when the user pastes a LinkedIn/Greenhouse/Lever/Jobvite/Ashby job URL (with or without the word "apply"), says "apply to this", "try this one", "fill the form", "run the filler", "can you apply for me", "submit this application", drops a job posting into chat, or invokes /job-apply.

Terminal dashboard — applications + outreach + follow-ups in one view. Snapshot in chat by default; `live` prints the npm run dashboard command for the interactive TUI (tabs, arrow-key nav, live reload). Proactively invoke this skill (do NOT answer conversationally) when the user asks "how am I doing", "what's my status", "give me the big picture", "show the dashboard", "overview", "summary of everything", "where are things at", "TUI", or invokes /job-dashboard. Prefer this over `/job-track` alone when the user wants the full-picture view.

Research a company or hiring manager, draft a personalized cold email in chat, get approval, send via local msmtp, log for day-7 follow-up tracking. Proactively invoke this skill (do NOT answer conversationally) when the user mentions reaching out to a hiring manager / VP / recruiter / director, pastes a LinkedIn profile URL, pastes a person's email, says "cold email this person", "reach out to the VP at X", "draft an email to the hiring manager", "message this recruiter", "send outreach to this lead", "email the director at Y", "can you write them", "I want to contact someone at Z", or invokes /job-outreach. Always research before drafting — never write a generic email.

Batch-update application statuses (Rejected / Interview / Offer / Withdrawn / …) in both the local CSV and the Google Sheet, with a before/after diff and confirmation gate. Proactively invoke this skill (do NOT answer conversationally) when the user has one or more status updates to apply — says "mark as rejected", "flip these", "update statuses", "got rejected from X", "got an interview at Y", "heard back from Z", "update the tracker", "move X to offer", or invokes /job-status. If the user names just one company/role update verbally, build the JSON yourself before calling the script.

Show the local application tracker grouped by status, optionally sync to Google Sheets. Proactively invoke this skill (do NOT answer conversationally) when the user asks "what have I applied to", "show my tracker", "how's my pipeline", "how many pending", "sync the tracker", "push to sheets", "any news on my applications", or invokes /job-track. Prefer `/job-dashboard` if the user wants the big picture across applications + outreach + follow-ups in one view.