Lessons & Meta — Living Knowledge Capture

The lessons system is a living record of local SEO observations that diverge from skill file expectations. Algorithm changes, GBP policy shifts, tool behaviors, and emerging AI search patterns get documented here when they contradict or extend existing skill guidance.

Purpose

Skills contain stable, well-established local SEO knowledge. But local search evolves — Google updates algorithms, GBP policies shift, AI search platforms change behavior, tools add or deprecate features.

The lessons file captures temporal observations that are true now but subject to change. Rather than editing skill files for every emerging pattern, lessons accumulate in a dedicated document that the agent references alongside skills.

Workflow

  1. Agent proposes — during a session, the agent encounters unexpected behavior that diverges from skill file guidance
  2. Human approves — the practitioner confirms the observation is valid and worth recording
  3. Agent documents — the approved lesson is added to the lessons file with structured metadata

No lesson is recorded without human approval. The agent proposes; the human validates.

Inclusion Criteria

Lessons belong here when they are:

  • Temporal — true now, subject to change
  • Exceptions to general skill file guidance in specific contexts
  • Emerging behaviors not yet captured in skill files

Lessons do NOT belong here when they are:

  • Already documented in existing skills
  • Single unverified data points
  • General best practices (those belong in relevant skill files)

Documentation Schema

Each lesson follows this format:

## [DATE] [CATEGORY]: [Title]
**Trigger:** Situation that surfaced the observation
**Discovery:** What was learned or what changed
**Confidence:** High | Medium | Low
**Scope:** Which skill areas are affected
**Expires:** Review date or "monitor"

Categories

CategoryWhat It Covers
ALGOAlgorithm changes affecting local rankings
GBP_POLICYGoogle Business Profile policy shifts
AI_SEARCHChanges in AI search platform behavior
TOOLTool API changes, deprecations, new capabilities
PATTERNRecurring patterns observed across multiple clients
WORKFLOWProcess improvements or workflow adjustments

Example Lesson

## 2026-03-15 GBP_POLICY: Service Area Display Changes
**Trigger:** Client's GBP stopped showing service area radius after update
**Discovery:** Google now displays service areas as a list of cities rather
  than a radius circle. This affects how service area pages should reference
  coverage — city lists are more accurate than radius claims.
**Confidence:** High (confirmed across 4 clients)
**Scope:** gbp-optimization, service-area-seo, local-landing-pages
**Expires:** monitor

Confidence Levels

  • High — confirmed across multiple clients or officially documented
  • Medium — observed in 2-3 instances, consistent pattern
  • Low — single observation, needs further validation

Low-confidence lessons get a shorter review date. If subsequent sessions don’t confirm the pattern, the lesson is removed.

Expiration and Review

Lessons include an expiration field — either a specific review date or “monitor” for ongoing observation. When a lesson reaches its review date:

  • If still valid: extend the expiration
  • If no longer valid: remove the lesson
  • If confirmed as permanent: migrate the knowledge into the relevant skill file

The lessons file stays current because stale observations get pruned.

Location

The lessons file lives at meta/lessons.md in the LocalSEOSkills repository. It’s version-controlled alongside skills but serves a different purpose — skills are stable knowledge, lessons are temporal observations.


Learn More

To learn how LocalSEOSkills works as a system, see how LocalSEOSkills works.