observing users

Capture user feedback as structured diagnostic observations using the Level 3 framework. Create or update individual use

ingesting dms

Import a Discord DM conversation export for a single user, resolve their wallet, pull their Score API profile, and produ

batch observing

Process multiple user DM exports in parallel using Claude Code native teams. A leader agent spawns one worker per user,

feedback observing

Read agent interaction logs across all packs, detect operational patterns (error rates, duration outliers, zero-invocati

concierge testing

Validate a gap hypothesis before any code is written. Manually simulate the proposed feature for ONE user whose canvas s

shaping journeys

Manage user canvases and shape common patterns into journey definitions for flow diagramming and testing.

daily synthesis

Automated feedback pipeline that runs daily (or on demand). Pulls new UI feedback from Supabase, enriches with Score API

shaping

Golden path command that consolidates journey patterns across canvases and files gap issues for product work.

level 3 diagnostic

This skill transforms support conversations into product intelligence. Instead of the traditional report → investigate →

analyzing gaps

Compare user expectations captured in UTCs and journeys with actual code behavior documented in reality files. Generate

detecting drift

Show what changed since the last validation of a specific artifact. Queries git history and feedback events to produce a

detecting staleness

Scan all artifacts with confidence metadata, compute freshness scores using a pure derived formula, and output a stalene

filing gaps

Create structured issues from Laboratory gap analysis reports with proper taxonomy labels and artifact linking.

batch filing gaps

Scan all observer canvases for `### GAP-*` sections, deduplicate similar gaps across users, route to the correct repo (`

generating followups

Generate Mom Test follow-up messages using per-user context isolation. Each user's follow-up is generated by an isolated

importing research

Bulk migrate legacy user research profiles to the User Truth Canvas (UTC) format with automatic JTBD classification and

refreshing artifacts

Re-validate an artifact by routing to the appropriate re-validation skill, then update confidence inputs (NOT the comput

snapshotting

Capture a point-in-time MER for a wallet. Produces a 4-layer snapshot: data state, visual screenshot, user perception, a

thinking

Analyze canvases that need attention, distill structured fears and steering targets into per-user cognition sidecar file

listening

Golden path command that auto-chains the Observer's intake pipeline. Ingests chronicle releases, daily synthesis feedbac

seeing

Golden path command that identifies stale canvases and refreshes their score API snapshots. Surfaces canvases missing ME

speaking

Golden path command that generates RLM-isolated follow-up messages with chronicle temporal context. Queries the chronicl

distilling

Analyze a single user's canvas state and produce structured cognition output. This skill runs as a **prompt-only subagen

growing

Operator-confirmed response attribution loop. Reviews proposed matches from `/listen`, classifies signal quality, update