Instagram follow audit adapter
A read-only relationship audit: diff a followers export against a following export and produce a citable report for unfollow decisions. It never unfollows, follows, or messages anyone.
Status: dev-only
Canonical sources: ig_follow_audit_helper.py and capture adapter contract
Status board#
| Capability | Status | Boundary |
|---|---|---|
| Inspect two user-provided export files | dev-only | Repository script; local files only |
| Preview the diff plan | dev-only | Counts only; no writes |
| Write the local audit report | dev-only | New empty report dir; report.md + report.json |
| Execute unfollows | parked | Never implemented here; reading never implies account action |
| Installed audit command | planned | Not included by the skill installer |
Commands#
python scripts/ig_follow_audit_helper.py inspect --followers followers.json --following following.json
python scripts/ig_follow_audit_helper.py preview --followers followers.json --following following.json
python scripts/ig_follow_audit_helper.py process --followers followers.json --following following.json --report-dir audit-reportEach side accepts an Instagram data-download JSON list or a plain text file with one handle per line. The user exports both lists from Instagram's own settings or data download; the helper has no network access and reads no browser state.
The report cites each suggested handle as [user N]. Acting on the list — unfollowing, keeping, messaging — is a separate human decision with its own consent; this adapter only produces evidence.