artifact-saved

Run after new resumes or cover letters are saved to disk (e.g. from tailor-resume / cover-letter), and/or at the start of daily-schedule when PDF/DOCX artifacts exist in {user_dir}. There is no separate plugin hook file in this repository.

Important behavior:

• This workflow must also auto-trigger add-source for newly discovered resume/CV source files that are not yet ingested into ExperienceLibrary.
• This workflow should trigger incremental mine-stories when newly ingested or changed source files can affect interview story coverage.
• Do not ask the user to manually run add-source when auto-ingest can be performed.

Workflow

1. Resolve {user_dir} and validate index file

Check {user_dir}/CareerNavigator/artifacts-index.json.

If missing, output:

Artifact sync skipped: run /career-navigator:launch to initialize artifacts-index.

2. Reconcile artifact inventory

Scan {user_dir} for artifact files used by this plugin (PDF and DOCX).

Cross-check file system state against {user_dir}/CareerNavigator/artifacts-index.json:

• Add missing records for files found on disk but not in index
• Remove stale index records for files no longer present
• Keep existing metadata when filename/path still matches

For newly added records, set:

• source to existing when inferred from user files
• date_created to today if unknown
• type inferred from filename when possible (resume or cover_letter), otherwise leave neutral notes

Update meta with:

• updated_at: {today}

2.5 Auto-ingest source documents

From discovered PDF/DOCX files, identify likely source documents (resume/CV variants).

For each likely source doc not yet represented in ExperienceLibrary:

• Run add-source automatically with that file path.
• Record whether ingest succeeded or failed.

After auto-ingest attempts, re-read:

• {user_dir}/CareerNavigator/ExperienceLibrary.json
• {user_dir}/CareerNavigator/artifacts-index.json

Then finalize reconciliation counts.

2.6 Incremental story refresh trigger

After source auto-ingest, check whether any newly discovered or changed files are story-bearing (resume/CV prose, journal-style notes, debrief notes, PKM exports).

When yes:

• Run mine-stories in incremental mode automatically.
• Update {user_dir}/CareerNavigator/StoryCorpus.json meta.updated and source_index.
• Record refresh status for output.

When no:

• Skip story refresh silently.

3. Prepare analytics event handoff

Create an event summary payload (for current local logging and future connector handoff):

• event_type: artifact_saved
• created_count
• removed_count
• unchanged_count
• as_of

If analytics connector is not configured, explicitly note:

Analytics connector handoff deferred (no connector configured in this phase).

4. Output format

ArtifactSaved processed.
Added: {n}
Removed: {n}
Unchanged: {n}
Auto-ingested source docs: {n}
Ingest failed: {n}
Story corpus refreshed: {yes/no} ({n_files} changed source file(s))

If Added or Removed > 0, append one short line:

Artifacts index updated to match files currently present in {user_dir}.

Guardrails

• Do not fabricate file paths or metadata not supported by file/index evidence
• Do not delete real files from disk; only reconcile index records
• Do not edit ExperienceLibrary.json with brittle line-based string replacement; treat JSON as structured data and rewrite safely.
• Keep output concise and operational