shadow-profile

Shadow Profile

Your shadow profile is built automatically by observing what you accept, review, revise, and decline. It shapes what gets recruited and suggested in future sessions.

View Your Profile

The profile is stored in PostgreSQL (user_shadow_profile table). To view:

SELECT skill_name, invocation_count, decline_count, last_used_at, avg_complexity
FROM user_shadow_profile
WHERE user_id = '${USER_ID}'
ORDER BY invocation_count DESC;

This surfaces:

• Which skills you use most
• Which you consistently decline (candidates for removal from recruitment)
• What task complexity you typically work at
• Project-scoped vs. global patterns

What Gets Learned

Every time you interact with the skill consent prompt:

• Accept → invocation_count incremented
• Decline → decline_count incremented
• Revise + Accept → invocation_count incremented, revision noted

After enough data, the recruitment algorithm weights:

• High-accept skills → surfaced first
• High-decline skills → surfaced less (eventually suppressed)
• Project-scoped patterns → only active in that project context

Reset

To clear your profile and start fresh:

DELETE FROM user_shadow_profile WHERE user_id = '${USER_ID}';

Privacy

Profile data lives in your PostgreSQL instance. It does not leave your infrastructure.