everything-gemini 2.5.0 → 2.7.0

package/skills/cve-remediation-orchestrator/SKILL.md

@@ -0,0 +1,111 @@
+---
+name: cve-remediation-orchestrator
+description: >-
+  Orchestrates defense-in-depth vulnerability remediation, virtual patching, kernel hardening, and service restart verification for critical CVEs without relying on exploit tools. Use when mitigating zero-days, evaluating container runtime isolation, or applying emergency WAF/gateway filters.
+---
+
+# CVE Remediation & Hardening Orchestrator (`cve-remediation-orchestrator`)
+
+An enterprise security skill that coordinates four-layer defensive mitigations for critical CVEs, neutralizing attack vectors at the architecture, kernel, and gateway levels without referencing or running exploit scripts.
+
+## Core Philosophy: Exploit-Agnostic Defense
+
+Rather than tracking fragile or dangerous exploit PoCs, this skill enforces **architectural invariants** that render entire vulnerability classes inoperable:
+1. **Virtual Patching**: Block malformed stream headers and buffer boundary exploits at the API gateway or WAF before they reach the runtime.
+2. **Kernel & Host Hardening**: Enforce Full Address Space Layout Randomization (ASLR), Data Execution Prevention (DEP/NX), and strict Seccomp profiles to thwart memory exploitation primitives.
+3. **Container Confinement**: Apply read-only root filesystems and drop all Linux capabilities to prevent post-exploitation persistence.
+4. **Permanent Remediation**: Verify vendor package upgrades and ensure web worker daemons (`php-fpm`, `nginx`, `apache2`) are restarted so no stale vulnerable inodes remain mapped in RAM.
+
+---
+
+## Defense-in-Depth Architecture
+
+```mermaid
+graph TD
+    CVE[Critical CVE / Zero-Day Disclosed] --> L1[Layer 1: Virtual Patching]
+    L1 --> L2[Layer 2: Kernel & ASLR Hardening]
+    L2 --> L3[Layer 3: Container Least Privilege]
+    L3 --> L4[Layer 4: Package Upgrade & Worker Restart]
+    
+    L1 -->|Gateway Regex / Header Filter| Drop1[Blocked at Perimeter]
+    L2 -->|sysctl randomize_va_space=2| Drop2[Exploit Memory Layout Broken]
+    L3 -->|readOnlyRootFilesystem| Drop3[Payload Write Denied]
+    L4 -->|Clean Inodes in RAM| Secured[Verified Secure]
+```
+
+---
+
+## Remediation Runbook
+
+### Layer 1: Ingress Virtual Patching (Envoy / WAF)
+When official runtime packages cannot be deployed immediately, apply an edge filter to drop malicious header boundaries:
+
+```yaml
+# Envoy Gateway Filter Example
+apiVersion: gateway.envoyproxy.io/v1alpha1
+kind: EnvoyFilter
+metadata:
+  name: drop-malformed-boundaries
+spec:
+  configPatches:
+    - applyTo: HTTP_FILTER
+      match:
+        context: SIDECAR_INBOUND
+      patch:
+        operation: INSERT_BEFORE
+        value:
+          name: envoy.filters.http.lua
+          typed_config:
+            "@type": type.googleapis.com/envoy.extensions.filters.http.lua.v3.Lua
+            inlineCode: |
+              function envoy_on_request(handle)
+                local content_type = handle:headers():get("content-type")
+                if content_type and string.find(content_type, "boundary=.*[\r\n\x00]") then
+                  handle:respond({[":status"] = "400"}, "Malformed stream boundary rejected.")
+                end
+              end
+```
+
+### Layer 2: Host & Kernel ASLR Hardening
+Ensure the host operating system enforces strict memory space randomization:
+
+```bash
+# Verify kernel ASLR level (2 = Full ASLR)
+cat /proc/sys/kernel/randomize_va_space
+
+# Enforce Full ASLR permanently
+sudo sysctl -w kernel.randomize_va_space=2
+echo "kernel.randomize_va_space = 2" | sudo tee -a /etc/sysctl.d/99-aslr-hardening.conf
+```
+
+### Layer 3: Container Runtime Confinement
+Harden Kubernetes deployment manifests against shellcode execution and disk persistence:
+
+```yaml
+securityContext:
+  allowPrivilegeEscalation: false
+  readOnlyRootFilesystem: true
+  runAsNonRoot: true
+  runAsUser: 10001
+  capabilities:
+    drop:
+      - ALL
+```
+
+### Layer 4: Permanent Patching & Worker Daemon Restart
+Verify that running worker processes are executing the upgraded package:
+
+```bash
+# Run automated defense verification
+bash skills/cve-remediation-orchestrator/scripts/verify-host-hardening.sh
+
+# Verify no running workers hold unlinked deleted executables
+ls -l /proc/*/exe 2>/dev/null | grep '(deleted)'
+```
+
+---
+
+## Related Skills
+- [phpcheck-server](file:///Users/salmanshaikh/development/github/everything-gemini/skills/phpcheck-server/SKILL.md) - Deep server and package backport auditor.
+- [security-audit](file:///Users/salmanshaikh/development/github/everything-gemini/skills/security-audit/SKILL.md) - General code, dependency, and vulnerability analysis.
+- [envoy-ai-gateway](file:///Users/salmanshaikh/development/github/everything-gemini/skills/envoy-ai-gateway/SKILL.md) - Production AI gateway routing, rate limiting, and virtual patching.

package/skills/cve-remediation-orchestrator/scripts/verify-host-hardening.sh

@@ -0,0 +1,72 @@
+#!/usr/bin/env bash
+
+# Host Hardening & Defense-in-Depth Verification Script
+# Verifies kernel ASLR, container capability constraints, and unlinked worker inodes in RAM.
+
+set -eo pipefail
+
+JSON_MODE=false
+if [[ "$1" == "--json" ]]; then
+  JSON_MODE=true
+fi
+
+# 1. Check Kernel ASLR
+ASLR_LEVEL="unknown"
+ASLR_STATUS="unknown"
+if [ -f /proc/sys/kernel/randomize_va_space ]; then
+  ASLR_LEVEL="$(cat /proc/sys/kernel/randomize_va_space)"
+  if [ "$ASLR_LEVEL" -eq 2 ]; then
+    ASLR_STATUS="hardened" # Full ASLR (stack, VDSO, mmap, brk)
+  elif [ "$ASLR_LEVEL" -eq 1 ]; then
+    ASLR_STATUS="partial"  # Partial ASLR
+  else
+    ASLR_STATUS="disabled" # Disabled ASLR (High Risk!)
+  fi
+elif command -v sysctl >/dev/null 2>&1; then
+  ASLR_VAL="$(sysctl -n vm.aslr 2>/dev/null || echo "unknown")"
+  if [ "$ASLR_VAL" != "unknown" ]; then
+    ASLR_LEVEL="$ASLR_VAL"
+    ASLR_STATUS="active"
+  else
+    ASLR_STATUS="unsupported_platform"
+  fi
+fi
+
+# 2. Check for unlinked/deleted running binaries (Stale worker trap)
+STALE_WORKERS_COUNT=0
+if [ -d /proc ]; then
+  STALE_WORKERS_COUNT="$(ls -l /proc/*/exe 2>/dev/null | grep -c '(deleted)' || true)"
+fi
+
+# 3. Overall Verdict
+VERDICT="pass"
+if [ "$ASLR_STATUS" = "disabled" ] || [ "$STALE_WORKERS_COUNT" -gt 0 ]; then
+  VERDICT="warn"
+fi
+
+if [ "$JSON_MODE" = true ]; then
+  cat <<EOF
+{
+  "defense_in_depth_verdict": "$VERDICT",
+  "kernel_aslr": {
+    "level": "$ASLR_LEVEL",
+    "status": "$ASLR_STATUS"
+  },
+  "stale_memory_workers_detected": $STALE_WORKERS_COUNT
+}
+EOF
+else
+  echo "🛡️ Host Hardening & Defense-in-Depth Verification"
+  echo "================================================"
+  echo "Kernel ASLR Status:  $ASLR_STATUS (level: $ASLR_LEVEL)"
+  echo "Stale Memory Workers: $STALE_WORKERS_COUNT deleted process handles"
+  if [ "$STALE_WORKERS_COUNT" -gt 0 ]; then
+    echo "⚠️ Warning: Daemons were updated on disk but not restarted in RAM."
+  fi
+fi
+
+if [ "$VERDICT" = "pass" ]; then
+  exit 0
+else
+  exit 1
+fi

package/skills/ebpf-microvm-sandbox/SKILL.md

@@ -0,0 +1,113 @@
+---
+name: ebpf-microvm-sandbox
+description: Enforces rootless ephemeral Micro-VM isolation, Linux Landlock filesystem jails, and Cilium eBPF network filtering for zero-trust speculative tool and code execution.
+---
+
+# eBPF Micro-VM Sandbox (`ebpf-microvm-sandbox`)
+
+Provides kernel-enforced rootless ephemeral Micro-VM isolation (Firecracker and gVisor), Linux Landlock path jails, seccomp-bpf syscall filtering, and Cilium eBPF network security for untrusted, speculative tool execution in Google Antigravity and Gemini agent loops.
+
+---
+
+## 1. Architectural Model: Kernel-Enforced Isolation Hierarchy
+
+```
+┌─────────────────────────────────────────────────────────────────────────────┐
+│                    ZERO-TRUST AGENT SANDBOX RUNTIME                         │
+│                                                                             │
+│   ┌─────────────────────────────────────────────────────────────────────┐   │
+│   │ 1. Antigravity Agent Runtime (Host Process)                         │   │
+│   │    • Speculative tool proposal (e.g. `run_command`, `eval_python`)  │   │
+│   │    • Intercepted by PreToolUse Lifecycle Security Hook              │   │
+│   └──────────────────────────────────┬──────────────────────────────────┘   │
+│                                      │                                      │
+│                                      ▼                                      │
+│   ┌─────────────────────────────────────────────────────────────────────┐   │
+│   │ 2. OPA Policy Gatekeeper & Sandbox Dispatcher                       │   │
+│   │    • Validates command safety signature and environment constraints │   │
+│   │    • Allocates ephemeral rootless micro-VM workspace                │   │
+│   └──────────────────────────────────┬──────────────────────────────────┘   │
+│                                      │                                      │
+│                                      ▼                                      │
+│   ┌─────────────────────────────────────────────────────────────────────┐   │
+│   │ 3. Ephemeral Micro-VM Jail (Firecracker / gVisor runsc)              │   │
+│   │    ┌────────────────────────────────────────────────────────────┐   │   │
+│   │    │ A. Filesystem: Linux Landlock V4 ephemeral copy-on-write   │   │   │
+│   │    │    • Host files outside sandbox mounted READ-ONLY          │   │   │
+│   │    │    • Temp workdir deleted on tool completion               │   │   │
+│   │    ├────────────────────────────────────────────────────────────┤   │   │
+│   │    │ B. Syscalls: Strict Seccomp-BPF whitelist filter           │   │   │
+│   │    │    • Blocks ptrace, mount, bpf, reboot, setuid, unshare    │   │   │
+│   │    ├────────────────────────────────────────────────────────────┤   │   │
+│   │    │ C. Network: Cilium eBPF Socket Filtering & Egress Policy   │   │   │
+│   │    │    • Strict CIDR whitelist (e.g. npmjs, pypi, googleapis)  │   │   │
+│   │    │    • Blocks RFC 1918 internal subnets & cloud metadata IPs │   │   │
+│   │    └────────────────────────────────────────────────────────────┘   │   │
+│   └──────────────────────────────────┬──────────────────────────────────┘   │
+│                                      │                                      │
+│                                      ▼                                      │
+│   ┌─────────────────────────────────────────────────────────────────────┐   │
+│   │ 4. Execution Telemetry & Differential Sync                          │   │
+│   │    • Captured stdout, stderr, exit code, and modified file diffs    │   │
+│   │    • Automatic rollback on policy violation or timeout abort        │   │
+│   └─────────────────────────────────────────────────────────────────────┘   │
+└─────────────────────────────────────────────────────────────────────────────┘
+```
+
+---
+
+## 2. Kernel Security Primitives
+
+### A. Linux Landlock V4 Path Rules
+- Speculative tools receive scoped path access:
+  - `LANDLOCK_ACCESS_FS_READ`: Allowed on workspace root.
+  - `LANDLOCK_ACCESS_FS_WRITE`: Confined strictly to `/tmp/agent-sandbox/<execution-id>`.
+  - Sensitive paths (`~/.ssh`, `~/.aws`, `~/.config/gcloud`, `/etc/shadow`) are masked with empty dummy inodes.
+
+### B. Seccomp-BPF Syscall Restriction
+- The sandbox enforces an allow-list BPF program at `execve`:
+  ```text
+  ALLOWED: read, write, openat, close, fstat, lseek, mmap, mprotect, munmap, brk, exit_group
+  DENIED (SIGKILL): ptrace, process_vm_readv, init_module, kexec_load, bpf, clone(NEWUSER)
+  ```
+
+### C. Cilium eBPF Network Quarantine
+- Egress is strictly blocked to Cloud Metadata Endpoints:
+  - `169.254.169.254/32` (AWS/GCP Instance Metadata Service)
+  - `10.0.0.0/8`, `172.16.0.0/12`, `192.168.0.0/16` (Internal LAN)
+- Only TLS port 443 egress to explicit DNS-validated dependency endpoints is permitted.
+
+---
+
+## 3. Operational Recipes
+
+### Launching an Isolated Speculative Command
+```bash
+# Verify sandbox status and host kernel security primitives
+everything-gemini doctor --profile full
+
+# Run speculative untrusted build in micro-VM jail
+everything-gemini resilience inject --target "speculative-build" --sandbox microvm
+```
+
+### Landlock Python Runner Pattern
+```python
+from google_antigravity.security import EphemeralMicroVMSandbox
+
+with EphemeralMicroVMSandbox(
+    memory_mb=512,
+    cpu_limit=1.0,
+    allow_network=False,
+    timeout_sec=30
+) as sandbox:
+    result = sandbox.exec(["python3", "-c", "import os; print(os.listdir('.'))"])
+    print(f"Sandbox Exit Code: {result.exit_code}")
+    print(f"Sandbox Output: {result.stdout}")
+```
+
+---
+
+## 4. Cross-Reference & Links
+- See [skills/ebpf-sandbox-enforcer/SKILL.md](file:///Users/salmanshaikh/development/github/everything-gemini/skills/ebpf-sandbox-enforcer/SKILL.md) for Landlock baseline rules.
+- See [skills/opa-gatekeeper-policy/SKILL.md](file:///Users/salmanshaikh/development/github/everything-gemini/skills/opa-gatekeeper-policy/SKILL.md) for OPA policy syntax.
+- See [agents/security-auditor.md](file:///Users/salmanshaikh/development/github/everything-gemini/agents/security-auditor.md) for automated vulnerability auditing.

package/skills/gemini-live-screen-vision/SKILL.md

@@ -0,0 +1,109 @@
+---
+name: gemini-live-screen-vision
+description: Streams high-framerate desktop/browser screen frames, analyzes visual DOM repaints, and extracts terminal OCR errors for live multimodal voice/visual pair programming with Gemini Live.
+---
+
+# Gemini Live Screen Vision & Repaint Analyzer (`gemini-live-screen-vision`)
+
+Integrates real-time desktop screen capture, browser DOM repaint diffing, terminal OCR error extraction, and synchronized audio-visual feedback loops directly with the Gemini 3.8 Live API for sub-second multimodal pair programming.
+
+---
+
+## 1. Architectural Model: Multimodal Video & Voice Duplex
+
+```
+┌─────────────────────────────────────────────────────────────────────────────┐
+│                 GEMINI LIVE SCREEN VISION STREAMING RUNTIME                 │
+│                                                                             │
+│   ┌─────────────────────────────────────────────────────────────────────┐   │
+│   │ 1. Client Screen & IDE Capture Engine                               │   │
+│   │    • 1080p @ 30–60fps WebRTC screen track or periodic JPEG stream   │   │
+│   │    • Active viewport region extraction (Browser canvas, Terminal)   │   │
+│   └──────────────────────────────────┬──────────────────────────────────┘   │
+│                                      │                                      │
+│                                      ▼                                      │
+│   ┌─────────────────────────────────────────────────────────────────────┐   │
+│   │ 2. Visual Repaint Diff & Change Detector                            │   │
+│   │    • Pixel-level perceptual hash (pHash) & structural similarity    │   │
+│   │    • Filters static frames; transmits only on repaint transitions   │   │
+│   └──────────────────────────────────┬──────────────────────────────────┘   │
+│                                      │                                      │
+│                                      ▼                                      │
+│   ┌─────────────────────────────────────────────────────────────────────┐   │
+│   │ 3. Gemini Live Bidirectional WebSocket Gateway                      │   │
+│   │    ┌───────────────────────────┐     ┌──────────────────────────┐   │   │
+│   │    │ Video Inbound: VP9/H.264  │     │ Audio Duplex: Opus 24kHz │   │   │
+│   │    │ (DOM renders, stack traces│     │ (Real-time voice pairs)  │   │   │
+│   │    └─────────────┬─────────────┘     └────────────┬─────────────┘   │   │
+│   └──────────────────┼────────────────────────────────┼─────────────────┘   │
+│                      └────────────────┬───────────────┘                     │
+│                                       ▼                                     │
+│   ┌─────────────────────────────────────────────────────────────────────┐   │
+│   │ 4. Multimodal Cognitive Fusion Engine                               │   │
+│   │    • Correlates developer speech ("why is the button misaligned?")   │   │
+│   │      with exact visual bounding box coordinates on active screen    │   │
+│   │    • Emits proactive vocal guidance before test failures compound   │   │
+│   └──────────────────────────────────┬──────────────────────────────────┘   │
+│                                      │                                      │
+│                                      ▼                                      │
+│   ┌─────────────────────────────────────────────────────────────────────┐   │
+│   │ 5. Surgical IDE Code Patching                                       │   │
+│   │    • Proposes atomic diffs directly to IDE active editor buffer     │   │
+│   └─────────────────────────────────────────────────────────────────────┘   │
+└─────────────────────────────────────────────────────────────────────────────┘
+```
+
+---
+
+## 2. Streaming Primitives & Frame Optimization
+
+### A. Perceptual Diff Filter (pHash)
+Streaming raw 60fps 1080p video incurs excessive bandwidth and token consumption. The vision engine computes frame-to-frame hamming distance:
+- $\Delta(\text{Frame}_t, \text{Frame}_{t-1}) < 5$: Static screen; transmission bypassed.
+- $\Delta \ge 5$: Visual transition (e.g. page repaint, terminal scroll); frame emitted to Gemini Live.
+
+### B. High-Precision Bounding Box Grounding
+When Gemini Live detects UI layout anomalies, it outputs normalized coordinate bounding boxes $[y_{\min}, x_{\min}, y_{\max}, x_{\max}]$:
+```json
+{
+  "anomaly": "CSS Flexbox Overflow in Hero Title",
+  "bounding_box": [120, 450, 240, 890],
+  "recommendation": "Set flex-wrap: wrap on #hero-container"
+}
+```
+
+---
+
+## 3. Operational Recipes
+
+### Launching Live Screen Vision Session via CLI
+```bash
+# Start live screen streaming bridge to Gemini Live API
+everything-gemini live-vision stream --fps 15 --resolution 1080p
+
+# Capture single-frame visual diagnostic of active IDE
+everything-gemini live-vision capture --workspace . --json
+```
+
+### Python SDK Multimodal Screen Session
+```python
+from google_antigravity.live import GeminiLiveVisionSession
+
+session = GeminiLiveVisionSession(
+    model="gemini-3.8-live",
+    audio_format="opus",
+    video_fps=15,
+    enable_vad=True
+)
+
+# Attach desktop window display capture
+session.attach_display_stream(display_id=1)
+session.start_bidirectional_loop()
+```
+
+---
+
+## 4. Cross-Reference Links
+- See [skills/gemini-live-multimodal/SKILL.md](file:///Users/salmanshaikh/development/github/everything-gemini/skills/gemini-live-multimodal/SKILL.md) for core WebSocket protocols.
+- See [skills/webrtc-agent-gateway/SKILL.md](file:///Users/salmanshaikh/development/github/everything-gemini/skills/webrtc-agent-gateway/SKILL.md) for WebRTC ICE negotiation.
+- See [agents/browser-qa.md](file:///Users/salmanshaikh/development/github/everything-gemini/agents/browser-qa.md) for automated visual QA assertions.

package/skills/genetic-prompt-cross-pollinator/SKILL.md

@@ -0,0 +1,110 @@
+---
+name: genetic-prompt-cross-pollinator
+description: Optimizes agent prompt surfaces and subagent contracts through multi-generation tournament selection, genetic crossover, and Pareto-frontier fitness evaluation.
+---
+
+# Genetic Prompt Cross-Pollinator (`genetic-prompt-cross-pollinator`)
+
+Implements Harness³ autonomous agentic Darwinism: applying evolutionary algorithms, multi-generation tournament selection, genetic crossover of prompt surfaces, and Pareto-frontier fitness evaluation to recursively discover optimal agent prompts without altering base model weights.
+
+---
+
+## 1. Evolutionary Architecture: Genetic Prompt Recombination
+
+```
+┌─────────────────────────────────────────────────────────────────────────────┐
+│                 GENETIC PROMPT CROSS-POLLINATION PIPELINE                   │
+│                                                                             │
+│   ┌─────────────────────────────────────────────────────────────────────┐   │
+│   │ 1. Seed Population Generation ($N=8$ Harness Candidates)            │   │
+│   │    • Baseline prompts from `.agents/`, `rules/`, and `skills/`      │   │
+│   │    • Stochastic prompt perturbations across 8 harness surfaces      │   │
+│   └──────────────────────────────────┬──────────────────────────────────┘   │
+│                                      │                                      │
+│                                      ▼                                      │
+│   ┌─────────────────────────────────────────────────────────────────────┐   │
+│   │ 2. Tournament Selection ($K=3$ Candidates per Bracket)              │   │
+│   │    • Multi-task probe evaluation on local test suite                │   │
+│   │    • Fitness function: $F = \frac{\text{PassRate} \times \text{Resilience}}{\text{Tokens} \times \text{Latency}}$ │
+│   └──────────────────────────────────┬──────────────────────────────────┘   │
+│                                      │                                      │
+│                                      ▼                                      │
+│   ┌─────────────────────────────────────────────────────────────────────┐   │
+│   │ 3. Genetic Crossover & Semantic Recombination                       │   │
+│   │    ┌───────────────────────────┐     ┌──────────────────────────┐   │   │
+│   │    │ Parent A: High Accuracy   │     │ Parent B: Low Token Burn │   │   │
+│   │    │ (Deep reasoning guidance) │     │ (Concise AST formatting) │   │   │
+│   │    └─────────────┬─────────────┘     └────────────┬─────────────┘   │   │
+│   └──────────────────┼────────────────────────────────┼─────────────────┘   │
+│                      └────────────────┬───────────────┘                     │
+│                                       ▼                                     │
+│   ┌─────────────────────────────────────────────────────────────────────┐   │
+│   │ 4. Mutation & Adversarial Validation                                │   │
+│   │    • Mutation Rate: $P_m = 0.15$ (Synonym swap, bullet compaction)  │   │
+│   │    • Adversarial fuzz verification against prompt regression probes │   │
+│   └──────────────────────────────────┬──────────────────────────────────┘   │
+│                                      │                                      │
+│                                      ▼                                      │
+│   ┌─────────────────────────────────────────────────────────────────────┐   │
+│   │ 5. Pareto Frontier Composition & Master Deployment                  │   │
+│   │    • Selects non-dominated solutions balancing cost vs accuracy     │   │
+│   │    • Commits winning harness phenotype to `.agents/` repository     │   │
+│   └─────────────────────────────────────────────────────────────────────┘   │
+└─────────────────────────────────────────────────────────────────────────────┘
+```
+
+---
+
+## 2. Core Mathematical Formalism
+
+### A. Multi-Objective Fitness Metric ($F$)
+Every candidate harness configuration $h_i$ is evaluated against an empirical test suite:
+$$F(h_i) = w_1 \cdot \text{Accuracy}(h_i) + w_2 \cdot \text{Resilience}(h_i) - w_3 \cdot \log(\text{Tokens}(h_i)) - w_4 \cdot \text{Latency}(h_i)$$
+
+Where:
+- $\text{Accuracy} \in [0, 1]$: Unit and integration test pass percentage.
+- $\text{Resilience} \in [0, 1]$: Survival rate under fault injection (Harness³ mutation testing).
+- $\text{Tokens}$: Cumulative prompt and output token consumption.
+- Weights $w = [0.45, 0.35, 0.10, 0.10]$ prioritizing correctness and stability.
+
+### B. Genetic Operators
+1. **Uniform Crossover**: Selects prompt sections (role instructions, tool guidelines, format envelopes) randomly from winning parent pairs.
+2. **Point Mutation**: Mutates prompt temperature, thinking token caps, and few-shot examples.
+3. **Elitism**: Top 10% highest-scoring harnesses pass unaltered into the next generation.
+
+---
+
+## 3. Operational Recipes
+
+### Running Genetic Evolution via CLI
+```bash
+# Evolve prompt surfaces across 5 generations with tournament selection
+everything-gemini genetic evolve --generations 5 --population 8 --max-usd 2.50
+
+# Inspect Pareto frontier trade-offs between cost and accuracy
+everything-gemini genetic pareto --workspace .
+```
+
+### Python SDK Integration Blueprint
+```python
+from google_antigravity.darwin import GeneticPromptOptimizer
+
+optimizer = GeneticPromptOptimizer(
+    target_skill="skills/cross-repo-swarm-sync/SKILL.md",
+    population_size=8,
+    generations=4,
+    tournament_k=3,
+    model="gemini-3.8-thinking"
+)
+
+winning_harness = optimizer.run_evolution()
+print(f"Optimal Fitness Score: {winning_harness.fitness_score:.4f}")
+winning_harness.commit_to_workspace()
+```
+
+---
+
+## 4. Cross-Reference Links
+- See [skills/harness3-mutation-testing/SKILL.md](file:///Users/salmanshaikh/development/github/everything-gemini/skills/harness3-mutation-testing/SKILL.md) for adversarial fault injection.
+- See [agents/darwin-prompt-engineer.md](file:///Users/salmanshaikh/development/github/everything-gemini/agents/darwin-prompt-engineer.md) for evolutionary subagent operations.
+- See [ROADMAP_AGI_v3.0.md](file:///Users/salmanshaikh/development/github/everything-gemini/ROADMAP_AGI_v3.0.md) for the Harness³ Darwinian roadmap.

package/skills/instinct-distillation-compiler/SKILL.md

@@ -0,0 +1,103 @@
+---
+name: instinct-distillation-compiler
+description: Compiles resolved edge cases, runtime exceptions, and developer corrections into enduring behavioral instinct rules without manual intervention.
+---
+
+# Learned Instinct Distillation & Rule Compiler (`instinct-distillation-compiler`)
+
+Continuously extracts, deduplicates, and compiles resolved edge cases, runtime stack traces, and developer feedback into persistent behavioral instinct rules within `.agents/hooks/learned-instincts/` to guarantee that errors resolved once are never repeated in future sessions.
+
+---
+
+## 1. Architectural Pipeline: Automated Instinct Distillation
+
+```
+┌─────────────────────────────────────────────────────────────────────────────┐
+│                 INSTINCT DISTILLATION & COMPILER PIPELINE                   │
+│                                                                             │
+│   ┌─────────────────────────────────────────────────────────────────────┐   │
+│   │ 1. Trajectory Exception & Correction Harvester                      │   │
+│   │    • Listens to PostToolUse failures and subsequent recovery turns  │   │
+│   │    • Captures developer corrections: "do not use X; use Y instead"  │   │
+│   └──────────────────────────────────┬──────────────────────────────────┘   │
+│                                      │                                      │
+│                                      ▼                                      │
+│   ┌─────────────────────────────────────────────────────────────────────┐   │
+│   │ 2. Semantic Distillation & Invariant Synthesis                      │   │
+│   │    • Strips ephemeral data (timestamps, temporary file paths)       │   │
+│   │    • Generalizes root cause: "When invoking tool T, ensure schema S"│   │
+│   └──────────────────────────────────┬──────────────────────────────────┘   │
+│                                      │                                      │
+│                                      ▼                                      │
+│   ┌─────────────────────────────────────────────────────────────────────┐   │
+│   │ 3. Deduplication & Conflict Resolution Engine                       │   │
+│   │    • Compares candidate rule against existing `.agents/rules/`      │   │
+│   │    • Semantic vector embedding match: if cosine similarity > 0.88,  │   │
+│   │      refines existing rule instead of appending duplicate          │   │
+│   └──────────────────────────────────┬──────────────────────────────────┘   │
+│                                      │                                      │
+│                                      ▼                                      │
+│   ┌─────────────────────────────────────────────────────────────────────┐   │
+│   │ 4. Deterministic Rule Compiler & Commit                             │   │
+│   │    • Formats valid markdown rule with clear triggers & constraints  │   │
+│   │    • Appends to `.agents/hooks/learned-instincts/rules/<topic>.md`  │   │
+│   └─────────────────────────────────────────────────────────────────────┘   │
+└─────────────────────────────────────────────────────────────────────────────┘
+```
+
+---
+
+## 2. Rule Structure & Compilation Grammar
+
+Compiled instinct rules strictly adhere to a standardized 3-part layout:
+```markdown
+# Instinct: [Descriptive Title]
+**Trigger Context**: [When the agent performs X on Y]
+**Observed Failure**: [Stack trace / root cause description]
+**Mandatory Invariant**:
+- DO: Always verify condition C before execution.
+- NEVER: Never invoke tool T with unvalidated parameter P.
+```
+
+### Example Compiled Rule:
+```markdown
+# Instinct: Safe BigQuery Partition Filtering
+**Trigger Context**: Running SQL queries against analytics tables with `run_command` (bq query).
+**Observed Failure**: Query failed due to missing partition filter scanning entire 10TB table.
+**Mandatory Invariant**:
+- DO: Always include `WHERE DATE(_PARTITIONTIME) >= DATE_SUB(CURRENT_DATE(), INTERVAL 7 DAY)`.
+- NEVER: Never run unrestricted `SELECT *` without explicit `LIMIT` and partition bounds.
+```
+
+---
+
+## 3. Operational Recipes
+
+### Triggering Instinct Compilation via CLI
+```bash
+# Review active learned instincts
+everything-gemini status --json | jq .instincts
+
+# Audit instinct repository and compile pending lessons
+everything-gemini audit --workspace .
+```
+
+### Python SDK Instinct Ingestion
+```python
+from google_antigravity.learning import InstinctDistiller
+
+distiller = InstinctDistiller(workspace_root=".")
+distiller.distill_correction(
+    trigger="npm test failed with missing module 'uuid'",
+    fix="Run npm install uuid --save-dev before running test suites",
+    category="dependency_hygiene"
+)
+distiller.commit_rules()
+```
+
+---
+
+## 4. Cross-Reference Links
+- See [skills/continuous-learning/SKILL.md](file:///Users/salmanshaikh/development/github/everything-gemini/skills/continuous-learning/SKILL.md) for session learning protocols.
+- See [hooks/learned-instincts/hooks.json](file:///Users/salmanshaikh/development/github/everything-gemini/hooks/learned-instincts/hooks.json) for the active instinct lifecycle hook.
+- See [agents/documentation-writer.md](file:///Users/salmanshaikh/development/github/everything-gemini/agents/documentation-writer.md) for rule curation standards.