All skills
simota avatar

/lore

@35ffd55
by shingo imotasimota/agent-skills85 stars
15

Curating cross-agent knowledge and institutional memory: extracts patterns from agent journals into METAPATTERNS.md, detects knowledge decay, propagates best practices. Use for memory curation.

Use this Skill: https://skilld.dev/gh/simota/agent-skills/lore

This session only. Nothing lands on disk.

referencepattern-taxonomy.md

≈1.4k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Pattern Taxonomy — Classification System

Purpose: Read this file when assigning pattern dimensions, generating IDs, structuring METAPATTERNS.md, or deciding lifecycle state transitions.

Contents

  • 4-dimension classification
  • Pattern ID convention
  • METAPATTERNS.md full schema
  • Pattern lifecycle

How Lore classifies, names, and structures patterns in METAPATTERNS.md.


4-Dimension Classification

Every pattern is tagged along 4 orthogonal dimensions:

Dimension 1: Domain

Domain Code Description Example Sources
INFRA Infrastructure, deployment, CI/CD, containers Gear, Scaffold, Gear[gha], Mend
APP Application logic, business rules, data models Builder, Schema, Gateway
TEST Testing strategies, coverage, quality assurance Radar, Voyager, Siege
DESIGN Architecture, patterns, code structure Atlas, Zen, Grove, Architect
PROCESS Workflows, collaboration, decision making Nexus, Sherpa, Magi, Judge
SECURITY Vulnerabilities, hardening, compliance Sentinel, Probe, Canon
PERF Performance, optimization, capacity Bolt, Tuner, Beacon
UX User experience, design, accessibility Palette, Flow, Prose, Echo
META Ecosystem-level patterns, agent interactions Darwin, Architect, Lore

Dimension 2: Type

Type Code Description Example
SUCCESS Proven approach that reliably works "Rolling restart resolves memory leaks in 90% of cases"
FAILURE Known approach that doesn't work in context "Increasing timeout masks upstream issues"
ANTI Common mistake to avoid "Alerting on causes instead of symptoms"
TRADEOFF Decision with inherent tension "Consistency vs availability in distributed cache"
HEURISTIC Rule of thumb, not absolute truth "If error rate > 2x baseline within 10min of deploy, rollback"

Dimension 3: Confidence

Level Code Evidence Count Trust Level
1 ANECDOTE 1 Low
2 EMERGING 2 Low-Medium
3-5 PATTERN 3-5 Medium
6-10 ESTABLISHED 6-10 High
11+ FOUNDATIONAL 11+ Very High

Dimension 4: Scope

Scope Code Description Propagation
AGENT Relevant to a single agent's domain Direct to that agent
CROSS Relevant across 2-5 agents Propagate to all relevant
ECOSYSTEM Relevant to the entire ecosystem Broadcast to Architect/Darwin/Nexus

Pattern ID Convention

[DOMAIN]-[TYPE]-[NNN]

Examples:
  INFRA-SUCCESS-001: "Rolling restart resolves gradual memory leaks"
  APP-ANTI-003: "Catching generic exceptions hides real errors"
  PROCESS-HEURISTIC-012: "Chain length > 5 agents correlates with lower success rate"
  META-TRADEOFF-002: "Agent specialization vs ecosystem complexity"

METAPATTERNS.md Full Schema

File Header

# METAPATTERNS — Ecosystem Knowledge Catalog

Last updated: [YYYY-MM-DD]
Total patterns: [N]
Confidence distribution: [N] Foundational / [N] Established / [N] Pattern / [N] Emerging / [N] Anecdote

---

Pattern Entry

## [PATTERN_ID]: [Title]

| Field | Value |
|-------|-------|
| **Domain** | [DOMAIN code] |
| **Type** | [TYPE code] |
| **Confidence** | [CONFIDENCE level] ([N] evidence) |
| **Scope** | [SCOPE code] |
| **Consumers** | [Agent1, Agent2, ...] |
| **Created** | [YYYY-MM-DD] |
| **Last validated** | [YYYY-MM-DD] |
| **Freshness** | [FRESH/CURRENT/AGING/STALE] |

**Pattern:** [Clear, actionable 1-3 sentence description]

**Evidence:**
1. [Agent] ([YYYY-MM-DD]): [What was observed, in what context]
2. [Agent] ([YYYY-MM-DD]): [What was observed, in what context]
3. [Agent] ([YYYY-MM-DD]): [What was observed, in what context]

**Implication:** [What consuming agents should do with this knowledge]

**Anti-pattern:** [What NOT to do, if applicable]

**Related:** [PATTERN_ID_1], [PATTERN_ID_2] (if any)

---

Index Section

At the end of METAPATTERNS.md, maintain an index:

## Index

### By Domain
- INFRA: [ID1], [ID2], ...
- APP: [ID3], [ID4], ...

### By Confidence (Highest First)
- FOUNDATIONAL: [ID1], [ID2], ...
- ESTABLISHED: [ID3], [ID4], ...

### By Freshness (Stalest First)
- STALE: [ID1], [ID2], ...
- AGING: [ID3], [ID4], ...

Pattern Lifecycle

CANDIDATE → REGISTERED → REINFORCED → PROMOTED → ... → AGING → ARCHIVED
                ↓                                         ↓
           CONTESTED → RESOLVED / SPLIT / SUPERSEDED    REMOVED
State Trigger Action
CANDIDATE New insight extracted, not yet validated Hold for additional evidence
REGISTERED First evidence + context documented Added to METAPATTERNS.md as Anecdote
REINFORCED Additional evidence adds to existing Update evidence list + confidence
PROMOTED Confidence threshold crossed Update level label
CONTESTED Contradictory evidence found Document both sides, seek resolution
AGING No new evidence for 90+ days Flag for review
ARCHIVED Superseded or no longer applicable Move to archive section
REMOVED Evidence invalidated Delete from catalog

Source: SKILL.md on GitHub

No alerts13d5 checks · Risk SAFE
  • Gen Agent Trust Hub13d

    Lore is a sophisticated knowledge curation and institutional memory skill designed to extract, synthesize, and manage patterns from agent journals and logs. It incorporates advanced defensive mechanisms against knowledge decay and information corruption, including confidence-based promotion rules and bi-temporal validity tracking, ensuring that the curated knowledge base remains accurate and actionable.

  • Socket13d

    No alerts

  • Snyk13d

    Risk: LOW · No issues

  • Runlayer6mo

    1/6 files flagged

  • ZeroLeaks5mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at 35ffd55. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 3 days ago.

Activeupdated 2 weeks ago

README badge

README badge for simota/agent-skills/lore