scoutica import

If you already keep your profile in an ai-job-search fork, import turns it into a Scoutica Skill Card in one step. It reads your fork’s local markdown, maps it onto the Scoutica schemas, and writes profile.json, rules.yaml, evidence.json, and SKILL.md.

Note

import is fully offline and deterministic. It never touches the network, never runs your shell, never calls an AI, and never commits to git. Every value from your fork is written through a structured serializer, so quotes, backslashes, or other special characters land as literal data — they can’t corrupt the output or execute. Evidence URLs are recorded verbatim and are never fetched.

Command Syntax

scoutica import aijs <fork-directory> [options]

The source type aijs selects the ai-job-search adapter. <fork-directory> is the path to your populated fork.

Options

Option	Behavior
--to <dir>	Where to write the card. Defaults to the current directory. Refuses a non-empty directory unless --force is given.
--salary-floor-eur <N>	Set your minimum base compensation (EUR) for each compensable engagement type. Must be a non-negative integer. Omitted if not provided — salary is never guessed from your CV.
--force	Allow writing into a non-empty target directory. Symlinked target files and symlinked parent directories are still refused.

What gets imported

Scoutica field	Comes from your fork
title, seniority, skills, primary_domains	01-candidate-profile.md (Identity, Professional Experience, Technical Skills)
years_experience, education, spoken_languages, certifications, specializations	profile + CLAUDE.md fallback
rules.yaml engagement / remote policy / blocked industries	employment status, constraints, and 04-job-evaluation.md deal-breakers
filters.stack_keywords	search-queries.md plus your top skills
evidence.json	GitHub / portfolio links and publications that carry a URL

Seniority is inferred from your role titles and years of experience. If it can’t be resolved and you’re not in an interactive terminal, the import aborts rather than guessing.

What is deliberately NOT imported

To keep your published card minimal, the importer never copies your behavioral profile, writing-style notes, interview stories (STAR answers), application tracker, or any salary data files. These are private working material with no home in a shareable card.

Exit codes

Code	Meaning
0	Success — card written.
2	A required field (title / seniority source / skills / domains) is still an unfilled [PLACEHOLDER]. Nothing is written.
3	A required enum (seniority) could not be resolved and the run is non-interactive. Nothing is written.
4	Target refusal — the directory is non-empty (without --force), or a target file or parent directory is a symlink.

Example

# Convert your fork into a card in ./my-card, setting a salary floor
scoutica import aijs ~/ai-job-search --to ./my-card --salary-floor-eur 85000

# Then validate and (when ready) publish
scoutica validate ./my-card
scoutica publish ./my-card

After importing, review the generated files — especially rules.yaml — before publishing. The importer is a fast first draft, not the final word on your policies.

Attribution

The ai-job-search project is an independent open-source workflow by Mads Lorentzen (MIT). Scoutica bridges to it; it does not vendor or replace it.