linkedin-applier

Run the autonomous LinkedIn applier loop for a query. Args: [text-id] [--dry-run] [--limit N]

Apply to a single LinkedIn job posting end-to-end given its URL. Reads profile, decides apply/skip, generates tailored CV, fills form (Easy Apply or external ATS), logs result to applied.jsonl. Use one dispatch per job. The orchestrator passes a structured prompt with job_url, query_id, profile_path, applied_log_path, and dry_run flag.

Drive LinkedIn's Easy Apply modal end-to-end for a single job posting. Use from the job-applier subagent when apply_type == "easy_apply".

Generate a short (150–300 word) cover letter tailored to a single job posting. Invoked by apply-easy-apply or fill-ats-form ONLY when the form has a cover-letter / "tell us why" textarea.

Generate a tailored CV in Markdown and convert it to PDF for upload to a specific job application. Use this skill from the job-applier subagent right before form upload, after decide-apply-or-skip has returned APPLY.

Compose an honest answer to a single screening question on a job application form, grounded in profile.md. Use from apply-easy-apply or fill-ats-form, once per question.

Decide whether the job-applier subagent should APPLY to or SKIP this job, based on profile + criteria + extracted job data. Returns one of {APPLY, SKIP} with a reason.