Skip to content

fix: add vetting phase to skill discovery to filter one-off activities#20

Open
steve-calvert-glean wants to merge 1 commit intomainfrom
fix/skill-discovery-vetting
Open

fix: add vetting phase to skill discovery to filter one-off activities#20
steve-calvert-glean wants to merge 1 commit intomainfrom
fix/skill-discovery-vetting

Conversation

@steve-calvert-glean
Copy link
Contributor

Summary

  • Adds a critical vetting phase to the skill discovery process
  • Filters out one-time events (team mergers, migrations, reorgs)
  • Prioritizes skills by cumulative value (frequency × time saved)
  • Shows vetting work including rejected candidates and reasons

Problem

The discover command was suggesting skills for activities like:

  • Team merger channel migrations (one-time organizational event)
  • Office move coordination (happens once every few years)

These aren't good skill candidates because they won't recur.

Solution

Added a structured vetting process that evaluates each candidate:

  1. Recurrence Test: Will this person do this again?
  2. Frequency Threshold: Daily > Weekly > Monthly > Reject
  3. Cumulative Value: frequency × time saved
  4. Disqualifying Signals: Explicit list of patterns to reject

The agent now outputs a "Rejected Candidates" section showing what was filtered and why.

Examples

Good candidates (include):

  • Daily standup prep (5 min × 5 days = 25 min/week)
  • PR review checklist (10 min × 10 PRs = 100 min/week)

Bad candidates (reject):

  • Team merger channel migration (2 hours, once)
  • Annual performance review (1 hour, once/year)

Test plan

  • Run /glean-skills:discover and verify it shows vetting work
  • Check that one-off activities are in "Rejected Candidates"
  • Verify recommendations focus on recurring patterns

🤖 Generated with Claude Code

The discover command was suggesting skills for one-time events like
team mergers and channel migrations. Added a critical vetting phase
that evaluates each candidate against:

- Recurrence test: Will this happen again?
- Frequency threshold: Daily/weekly/monthly prioritization
- Cumulative value: frequency × time saved
- Disqualifying signals: mergers, migrations, reorgs

The agent now shows its vetting work including rejected candidates
and reasons. Quality over quantity - a few high-value recommendations
beat many weak ones.

Co-Authored-By: Claude <noreply@anthropic.com>
@steve-calvert-glean steve-calvert-glean requested a review from a team as a code owner February 14, 2026 02:00
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant