Cognitive Flexibility Release

Cognitive Flexibility Skill - AI cognitive flexibility with 4 modes. Supports automatic mode switching and metacognitive monitoring. Use when: - Complex reas...

Security Scan

The skill's code, docs, and runtime instructions are internally consistent with its stated purpose (four cognitive modes, mode switching, self-assessment); it requires no external credentials or install steps and mainly reads/writes local logs — a few minor tool/metadata notes are worth reviewing before install.

Purpose & Capability

The name/description (cognitive flexibility with OOA/OODA/OOCA/OOHA modes) match the included implementation files (reasoner, controller, pattern matcher, self-assessor, usage monitor). There are no unrelated binaries, environment variables, or cloud credentials requested. The allowed tools (memory_search, web_search, Read/Write/Edit) are reasonable for a reasoning/monitoring skill that can consult memory and local files; sessions_send is listed but not used in the shown code.

Instruction Scope

SKILL.md and code instruct the agent to use tools (memory_search, web_search, Read/Write/Edit) and to log usage to local files. The code reads/writes only local files inside the skill (logs/, feedback files) and calls tools.memory_search when provided. There are no instructions to read arbitrary system files or to exfiltrate data to external endpoints. However: the allowed-tools list includes sessions_send and web_search (web_search is referenced in docs as an optional external verification step). If an agent is granted web_search or sessions_send at runtime those tools could transmit context externally — this is a capability the user should choose deliberately.

Install Mechanism

No install specification is provided (instruction-only for the platform) and there are no downloads or third‑party package installs in the manifest. The package includes only Python source that uses standard library modules. This is low-risk from an installation/extraction perspective.

Credentials

The skill declares no required environment variables, no primary credential, and no config paths. Code does not attempt to read process environment or secret files. This is proportionate to the skill's purpose.

Persistence & Privilege

always:false (default) and the skill does not request system-wide persistence. It writes diagnostic and usage logs under the skill's directory (logs/, feedback files) which is normal for monitoring. Autonomous invocation (model invocation enabled) is the platform default and not a unique privilege here.

Assessment

This skill appears coherent and implements the four cognitive modes as described. Before installing: 1) Confirm the skill source/owner (top-level metadata shows no homepage while the package files reference GitHub/ClawHub URLs); 2) Run the included tests locally (python tests/test_cognitive_skills.py) in a sandbox to validate behavior; 3) Note that the skill will create local logs/feedback files under its directory — review those files for any sensitive data your agent might write; 4) When granting runtime tools to the agent, be intentional: memory_search and local Read/Write are necessary for functionality, but web_search and sessions_send can cause data to be transmitted externally — disable them unless you need external lookups or inter-session messaging; 5) If you plan to publish or run the provided publish scripts, follow token/browser auth guidance carefully and avoid pasting sensitive tokens into logs or public places. If you want extra assurance, inspect the omitted files/tests for network calls or unexpected file system paths before enabling the skill for production workloads.

Cognitive Flexibility Skill

Overview

This Skill implements four cognitive modes based on human cognitive science:

Mode	Name	Driver	Scenario	Core Ability
OOA	Experience Mode	Memory-driven	Familiar scenarios	Pattern matching
OODA	Reasoning Mode	Knowledge-driven	Complex problems	Chain reasoning
OOCA	Creative Mode	Association-driven	Innovation needs	Analogy generation
OOHA	Discovery Mode	Hypothesis-driven	Exploration	Hypothesis generation

Quick Start

Basic Usage

from scripts.cognitive_controller import CognitiveController

# Create controller
controller = CognitiveController(confidence_threshold=0.7)

# Execute task (auto mode selection)
task = "Analyze user feedback data"
result = await controller.process(task, tools=tools)

# View result
print(f"Mode: {result['mode']}")
print(f"Answer: {result['answer']}")
print(f"Confidence: {result['assessment']['overall_score']:.2f}")

Manual Mode Selection

# OODA reasoning mode
from scripts.chain_reasoner import OODAReasoner
reasoner = OODAReasoner()
result = await reasoner.process(task, tools=tools)

# OOA experience mode
from scripts.pattern_matcher import PatternMatcher
matcher = PatternMatcher()
result = await matcher.match(task, tools=tools)

# OOCA creative mode
from scripts.creative_explorer import CreativeExplorer
explorer = CreativeExplorer()
result = await explorer.explore(task)

# OOHA discovery mode
from scripts.hypothesis_generator import HypothesisGenerator
generator = HypothesisGenerator()
result = await generator.discover(task)

Features

• 4 Cognitive Modes: OOA/OODA/OOCA/OOHA
• Auto Mode Switching: Cognitive Controller selects best mode
• Metacognitive Monitoring: Self-assessment and confidence scoring
• Usage Tracking: Complete usage logs and statistics
• 100% Test Coverage: All tests passing

File Structure

cognitive-flexibility/
├── scripts/
│   ├── __init__.py
│   ├── chain_reasoner.py       # OODA reasoning
│   ├── pattern_matcher.py      # OOA pattern matching
│   ├── self_assessor.py        # Metacognitive monitoring
│   ├── cognitive_controller.py # Mode switching
│   ├── creative_explorer.py    # OOCA creative mode
│   ├── hypothesis_generator.py # OOHA discovery mode
│   └── usage_monitor.py        # Usage tracking
├── references/
│   └── ooda-guide.md
├── tests/
│   └── test_cognitive_skills.py
├── SKILL.md
├── README.md
└── MONITORING-GUIDE.md

Testing

# Run tests
python tests/test_cognitive_skills.py

# Expected output: 6/6 tests passed (100%)

Monitoring

from scripts.usage_monitor import UsageMonitor

monitor = UsageMonitor()

# Get usage stats
stats = monitor.get_stats(days=7)

# Generate report
report = monitor.generate_report(days=7)
print(report)

Requirements

• Python >= 3.8
• OpenClaw >= 2026.3.28
• No external dependencies

License

MIT License

Support

• Documentation: See README.md and MONITORING-GUIDE.md
• Issues: GitHub Issues
• Community: Discord #skills-feedback

---

DaoShi · Cognitive Flexibility Skill v2.1.0