Profile settings

Profile is the foundation every agent reads. Get this right early; the agents tail-chase if it’s wrong.

Target role

A short phrase like “Senior Backend Engineer” or “Product Designer”. The job-searcher uses it as the primary query term + the title-match filter. The cv-adapter uses it to pick which experience bullets to elevate.

Suggestion: stay specific but not too narrow. “Senior” prefix matters — the agent will not apply to senior roles unless the profile says so.

Work preferences

• Job types — full-time, contract, freelance.
• Work arrangements — remote, hybrid, onsite.
• Remote regions — for remote-friendly only: which timezone bands.
• Onsite locations — city + country pairs, or “anywhere in EU”.

The job-searcher bakes these into the search query (we filter at source — post-hoc filtering on a too-broad query rejects results we already paid for).

Languages

ISO codes. Used by the cover-letter generation when a job posting is in a non-English language and the agent needs to know which languages it should output.

Document format

Default cv_format_mode and cl_format_mode for new applications. The Stage 3 picker overrides per-app — change the default here when you’ve decided your portfolio target ATS-portal-heavy companies that always need the parser-safe path. Detail in ats-mode.

Why agents read this every run

The MCP tool getAgentContext pulls profile + assets + writing style + job profile in one round-trip. Every cv-adapter, employer-profiler, and qa-agent run starts with that call. If the row is wrong, the next 50 generations will be wrong. Re-edit and re-run; the change takes effect immediately on the next agent invocation.