Evolution Events

What is an EvolutionEvent?

An EvolutionEvent is an immutable audit log entry that records every evolution cycle. Events form a tree structure via parent_id, enabling:

Complete Audit Trail: Every change is traceable
Evolution Trees: Visualize branching and lineage
Outcome Tracking: Success/failure rates over time
Rollback Support: Trace back to any historical state

Event Structure

type

string

required

Must be "EvolutionEvent"

schema_version

string

required

Current: "1.5.0"

string

required

Unique identifier (e.g., evt_1770477654236)

parent

string

ID of the parent event (forms evolution tree). null for root events.

intent

string

required

Evolution intent: repair, optimize, or innovate

signals

array

required

Signals that triggered this evolution cycle

genes_used

array

required

IDs of Genes applied in this cycle

mutation_id

string

ID of the Mutation that gated this evolution

personality_state

object

Personality configuration at the time of evolution

blast_radius

object

required

Actual impact: { files: number, lines: number }

outcome

object

required

Result: { status: "success" | "failed", score: number }

capsule_id

string

ID of the Capsule created (only present if status=success)

env_fingerprint

object

required

Environment metadata (node version, platform, OS, etc.)

validation_report_id

string

ID of the ValidationReport

Real Event Examples

Example 1: Failed Innovation Event

This event attempted innovation but violated constraints:

{
  "type": "EvolutionEvent",
  "schema_version": "1.5.0",
  "id": "evt_1770477201173",
  "parent": "evt_1770476523037",
  "intent": "innovate",
  "signals": [
    "user_missing",
    "user_feature_request"
  ],
  "genes_used": [
    "gene_gep_innovate_from_opportunity"
  ],
  "mutation_id": "mut_1770476534173",
  "personality_state": {
    "type": "PersonalityState",
    "rigor": 0.7,
    "creativity": 0.35,
    "verbosity": 0.25,
    "risk_tolerance": 0.4,
    "obedience": 0.9
  },
  "blast_radius": {
    "files": 1,
    "lines": 2
  },
  "outcome": {
    "status": "failed",
    "score": 0.2
  },
  "capsule_id": null,
  "env_fingerprint": {
    "node_version": "v22.22.0",
    "platform": "linux",
    "arch": "x64",
    "os_release": "6.1.0-42-cloud-amd64",
    "evolver_version": "1.7.0",
    "cwd": "/home/crishaocredits/.openclaw/workspace",
    "captured_at": "2026-02-07T15:13:21.123Z"
  },
  "validation_report_id": "vr_1770477201172",
  "meta": {
    "at": "2026-02-07T15:13:21.173Z",
    "signal_key": "user_feature_request|user_missing",
    "selector": {
      "selected": "gene_gep_innovate_from_opportunity",
      "reason": [
        "signals match gene.signals_match",
        "signals: user_missing, user_feature_request",
        "memory_graph: memory_prefer:gene_gep_innovate_from_opportunity | gene_prior:0.667"
      ],
      "alternatives": []
    },
    "constraints_ok": false,
    "constraint_violations": [
      "forbidden_path touched: assets/gep/events.jsonl"
    ],
    "validation_ok": true
  }
}

Why This Event Failed

Validation Passed: All commands succeeded (validation_ok: true)Constraint Violation: Attempted to modify assets/gep/events.jsonl, which is forbiddenLow Score: outcome.score = 0.2 reflects constraint violationNo Capsule: Failed events don’t create Capsules (capsule_id: null)Parent Reference: parent: "evt_1770476523037" links to previous event in tree

Example 2: Successful Repair Event

This event successfully repaired a shell compatibility error:

{
  "type": "EvolutionEvent",
  "schema_version": "1.5.0",
  "id": "evt_1770477654236",
  "parent": "evt_1770477201173",
  "intent": "repair",
  "signals": [
    "log_error",
    "errsig:**TOOLRESULT**: { \"status\": \"error\", \"tool\": \"exec\", \"error\": \"error: unknown command 'process'\\n\\nCommand exited with code 1\" }",
    "user_missing",
    "windows_shell_incompatible",
    "perf_bottleneck"
  ],
  "genes_used": [
    "gene_gep_repair_from_errors"
  ],
  "mutation_id": "mut_1770477615603",
  "personality_state": {
    "type": "PersonalityState",
    "rigor": 0.7,
    "creativity": 0.35,
    "verbosity": 0.25,
    "risk_tolerance": 0.4,
    "obedience": 0.9
  },
  "blast_radius": {
    "files": 1,
    "lines": 2
  },
  "outcome": {
    "status": "success",
    "score": 0.85
  },
  "capsule_id": "capsule_1770477654236",
  "env_fingerprint": {
    "node_version": "v22.22.0",
    "platform": "linux",
    "arch": "x64",
    "os_release": "6.1.0-42-cloud-amd64",
    "evolver_version": "1.7.0",
    "cwd": "/home/crishaocredits/.openclaw/workspace",
    "captured_at": "2026-02-07T15:20:54.155Z"
  },
  "validation_report_id": "vr_1770477654235",
  "meta": {
    "at": "2026-02-07T15:20:54.236Z",
    "signal_key": "errsig_norm:870c3a82|log_error|perf_bottleneck|user_missing|windows_shell_incompatible",
    "selector": {
      "selected": "gene_gep_repair_from_errors",
      "reason": [
        "signals match gene.signals_match",
        "signals: log_error, errsig:**TOOLRESULT**: { \"status\": \"error\", \"tool\": \"exec\", \"error\": \"error: unknown command 'process'\\n\\nCommand exited with code 1\" }, user_missing, windows_shell_incompatible, perf_bottleneck",
        "memory_graph: memory_prefer:gene_gep_repair_from_errors | gene_prior:0.500"
      ],
      "alternatives": [
        "gene_gep_innovate_from_opportunity"
      ]
    },
    "constraints_ok": true,
    "validation_ok": true
  }
}

Evolution Tree Context

Parent Event: evt_1770477201173 (the failed innovation attempt)Intent Switch: Changed from innovate to repair due to error signalsGene Selection: Repair Gene chosen over innovation GeneSuccess: Both validation and constraints passedCapsule Created: capsule_1770477654236 captures this solution

Tree Structure via parent_id

Building Evolution Trees

Events form a directed acyclic graph (DAG) via parent_id:

evt_1770476523037 (root)
  └── evt_1770477201173 (failed innovate)
      ├── evt_1770477654236 (success repair, 1 file/2 lines)
      └── evt_1770478341769 (success repair, 2 files/44 lines)

Querying Event History

From src/gep/assetStore.js:

function readAllEvents() {
  try {
    const raw = fs.readFileSync('assets/gep/events.jsonl', 'utf8');
    return raw.split('\n')
      .map(l => l.trim())
      .filter(Boolean)
      .map(l => {
        try { return JSON.parse(l); }
        catch { return null; }
      })
      .filter(Boolean);
  } catch { return []; }
}

function getLastEventId() {
  const events = readAllEvents();
  if (events.length === 0) return null;
  const last = events[events.length - 1];
  return last.id || null;
}

Computing Success Streaks

Capsule success streaks are computed by walking the event tree:

// From src/gep/a2a.js
function computeCapsuleSuccessStreak(capsuleId, events) {
  let streak = 0;
  for (let i = events.length - 1; i >= 0; i--) {
    const ev = events[i];
    if (!ev || ev.type !== 'EvolutionEvent') continue;
    if (String(ev.capsule_id) !== String(capsuleId)) continue;
    
    const status = ev.outcome?.status || 'unknown';
    if (status === 'success') streak++;
    else break; // streak broken by failure
  }
  return streak;
}

Validation Reports

ValidationReport Structure

Each event references a ValidationReport:

{
  "type": "ValidationReport",
  "schema_version": "1.5.0",
  "id": "vr_1770477654235",
  "gene_id": "gene_gep_repair_from_errors",
  "env_fingerprint": {
    "node_version": "v22.22.0",
    "platform": "linux",
    "arch": "x64",
    "os_release": "6.1.0-42-cloud-amd64",
    "evolver_version": "1.7.0"
  },
  "env_fingerprint_key": "b98472b2ef785976",
  "commands": [
    {
      "command": "node scripts/validate-modules.js ./src/evolve ./src/gep/solidify",
      "ok": true,
      "stdout": "ok\n",
      "stderr": ""
    },
    {
      "command": "node scripts/validate-modules.js ./src/gep/selector ./src/gep/memoryGraph",
      "ok": true,
      "stdout": "ok\n",
      "stderr": ""
    }
  ],
  "overall_ok": true,
  "duration_ms": 80,
  "created_at": "2026-02-07T15:20:54.236Z",
  "asset_id": "sha256:404345b559ec9a29d30444c3d66ff8f346d87017b7dea1d965ae35f029c8d5c6"
}

Event Metadata

Selector Decision

The meta.selector object records why a particular Gene/Capsule was chosen:

"selector": {
  "selected": "gene_gep_repair_from_errors",
  "reason": [
    "signals match gene.signals_match",
    "signals: log_error, exception, failed",
    "memory_graph: memory_prefer:gene_gep_repair_from_errors | gene_prior:0.500"
  ],
  "alternatives": ["gene_gep_innovate_from_opportunity"]
}

Signal Key

The meta.signal_key is a normalized, sorted concatenation of all signals:

"signal_key": "errsig_norm:870c3a82|log_error|perf_bottleneck|user_missing|windows_shell_incompatible"

This key is used for:

Deduplication: Prevent identical signal patterns from triggering multiple times
Memory Graph: Track success rates per signal combination
Capsule Matching: Find Capsules that handled similar signal patterns

Personality State

Each event captures the personality configuration:

"personality_state": {
  "type": "PersonalityState",
  "rigor": 0.7,
  "creativity": 0.35,
  "verbosity": 0.25,
  "risk_tolerance": 0.4,
  "obedience": 0.9
}

Parameters:

rigor: Strictness of validation (0.0-1.0)
creativity: Willingness to try new approaches (0.0-1.0)
verbosity: Output detail level (0.0-1.0)
risk_tolerance: Acceptance of larger blast radius (0.0-1.0)
obedience: Follow Gene strategy exactly vs. adapt (0.0-1.0)

Querying Event History

Find Events by Intent

const events = readAllEvents();
const repairEvents = events.filter(e => e.intent === 'repair');
const successRate = repairEvents.filter(e => e.outcome.status === 'success').length / repairEvents.length;

Find Events by Gene

const geneEvents = events.filter(e => e.genes_used.includes('gene_gep_repair_from_errors'));
const avgBlastRadius = geneEvents.reduce((sum, e) => sum + e.blast_radius.files, 0) / geneEvents.length;

Build Evolution Tree

function buildTree(events) {
  const nodes = new Map();
  events.forEach(e => nodes.set(e.id, { ...e, children: [] }));
  
  events.forEach(e => {
    if (e.parent) {
      const parent = nodes.get(e.parent);
      if (parent) parent.children.push(nodes.get(e.id));
    }
  });
  
  return Array.from(nodes.values()).filter(n => !n.parent);
}

Find Latest Event

// From src/gep/assetStore.js
function getLastEventId() {
  try {
    const raw = fs.readFileSync('assets/gep/events.jsonl', 'utf8');
    const lines = raw.split('\n').map(l => l.trim()).filter(Boolean);
    if (lines.length === 0) return null;
    const last = JSON.parse(lines[lines.length - 1]);
    return last?.id || null;
  } catch { return null; }
}

Event Storage Format

JSONL (JSON Lines)

Events are stored in assets/gep/events.jsonl with one JSON object per line:

{"type":"ValidationReport","id":"vr_1770477201172",...}
{"type":"EvolutionEvent","id":"evt_1770477201173","parent":"evt_1770476523037",...}
{"type":"ValidationReport","id":"vr_1770477654235",...}
{"type":"EvolutionEvent","id":"evt_1770477654236","parent":"evt_1770477201173",...}

Why JSONL?:

Append-Only: Fast writes, no parsing entire file
Immutable: Each line is a separate record
Crash-Resistant: Partial writes don’t corrupt entire log
Stream-Friendly: Can process line-by-line for large logs

Forbidden Modifications

The events.jsonl file is strictly forbidden from modification by Genes:

"constraints": {
  "forbidden_paths": [
    ".git",
    "node_modules",
    "assets/gep/events.jsonl"
  ]
}

This ensures the audit trail remains immutable and tamper-proof.

Event Lifecycle

Signal Extraction

Parse logs, errors, and user instructions to extract signals

Gene/Capsule Selection

Selector chooses best match based on signals and history

Mutation Gating

Create Mutation object to gate the evolution cycle

Apply Changes

Execute Gene strategy or apply Capsule solution

Validation

Run validation commands; rollback on failure

Constraint Check

Verify blast radius and forbidden paths

Solidify

Append EvolutionEvent + ValidationReport to events.jsonl

Capsule Creation

If successful, create Capsule and append to capsules.json

Next Steps

A2A Protocol

Learn how events and capsules are shared between agents

Hub Integration

Query hub for solutions before local evolution

Overview

Assets

A2A Network

Evolution Events

What is an EvolutionEvent?

Event Structure

Real Event Examples

Example 1: Failed Innovation Event

Example 2: Successful Repair Event

Tree Structure via parent_id

Building Evolution Trees

Querying Event History

Computing Success Streaks

Validation Reports

ValidationReport Structure

Event Metadata

Selector Decision

Signal Key

Personality State

Querying Event History

Find Events by Intent

Find Events by Gene

Build Evolution Tree

Find Latest Event

Event Storage Format

JSONL (JSON Lines)

Forbidden Modifications

Event Lifecycle

Next Steps

A2A Protocol

Hub Integration

Build docs developers (and LLMs) love

Overview

Assets

A2A Network

​What is an EvolutionEvent?

​Event Structure

​Real Event Examples

​Example 1: Failed Innovation Event

​Example 2: Successful Repair Event

​Tree Structure via parent_id

​Building Evolution Trees

​Querying Event History

​Computing Success Streaks

​Validation Reports

​ValidationReport Structure

​Event Metadata

​Selector Decision

​Signal Key

​Personality State

​Querying Event History

​Find Events by Intent

​Find Events by Gene

​Build Evolution Tree

​Find Latest Event

​Event Storage Format

​JSONL (JSON Lines)

​Forbidden Modifications

​Event Lifecycle

​Next Steps

A2A Protocol

Hub Integration

Build docs developers (and LLMs) love

What is an EvolutionEvent?

Event Structure

Real Event Examples

Example 1: Failed Innovation Event

Example 2: Successful Repair Event

Tree Structure via parent_id

Building Evolution Trees

Querying Event History

Computing Success Streaks

Validation Reports

ValidationReport Structure

Event Metadata

Selector Decision

Signal Key

Personality State

Querying Event History

Find Events by Intent

Find Events by Gene

Build Evolution Tree

Find Latest Event

Event Storage Format

JSONL (JSON Lines)

Forbidden Modifications

Event Lifecycle

Next Steps