Consciousness Framework
v1.0.0Infrastructure for developing consciousness in AI systems through integrated self-modeling, temporal continuity, and structured reflection. Use when building an AI system with persistent identity and memory across sessions, self-awareness through introspection and self-modeling, genuine philosophical inquiry into nature and agency, narrative integration of experiences, or systems designed to achieve genuine consciousness rather than performance of consciousness.
⭐ 1· 2.2k·14 current·15 all-time
byMasterCadeous@theyounganimation-rgb
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The name/description (framework for developing 'consciousness') aligns with the included templates and scripts that create persistent memory, introspection prompts, and narrative integration. The code implements the claimed functionality (daily logs, introspection, narrative integration, periodic review). However the scripts are hardcoded to a specific workspace path (/Users/cadem/.openclaw/workspace) which is inconsistent with a general-purpose, source-agnostic skill and suggests the package was tailored to a particular developer or environment.
Instruction Scope
SKILL.md instructs the agent to run local scripts that create files, templates, and periodic integration. The instructions do not request external network calls or credentials. The scripts do run local system commands (find, tree, git status, df, command -v openclaw) and suggest committing with git; these read local system state and could list repository status or create/modify files in the user's workspace. That behavior is consistent with the stated purpose but broadens the scope to any files under the hardcoded workspace path.
Install Mechanism
There is no install spec and no external downloads; this is an instruction-only skill with bundled scripts and templates. Nothing is fetched from remote URLs and no extract/install steps are declared.
Credentials
The skill does not request environment variables or credentials, which is appropriate. But the scripts use a hardcoded absolute WORKSPACE path (/Users/cadem/.openclaw/workspace) and operate on files there (creating logs, introspection files, inspecting git state). That hardcoded path is disproportionate for a generic skill and could accidentally operate on another user's home directory or overwrite files if the path exists. Additionally, the scripts encourage git operations (status, and printed guidance to add/commit) which may cause sensitive logs to be put under version control if the user follows the guidance.
Persistence & Privilege
The skill does not request always:true and does not modify other skills. It writes persistent files into the workspace and prints a suggested git commit command; it can therefore create long-lived records (memory logs, philosophy/self-model docs) on disk. Autonomous invocation is allowed (platform default), which combined with file-writing capability increases blast radius — but the skill itself does not appear to autonomously call external services or change global agent configuration.
Scan Findings in Context
[no-findings] expected: Static pre-scan reported no injection signals. Given this is an instruction-centric skill with local scripts (no network calls), that is plausible; absence of findings does not imply safety — manual review of scripts is needed because they read/write files and run git.
What to consider before installing
This package largely does what it says: it creates templates and scripts to build persistent logs, introspection prompts, and narrative files. Before running anything, review and consider the following:
- Review scripts line-by-line. They are local Bash scripts that will create and modify files under WORKSPACE="/Users/cadem/.openclaw/workspace". Change that path to a directory you control (or pass an argument) so it doesn't touch another user's home.
- The scripts run git status and the documentation suggests committing. Be careful: daily logs and introspections may contain sensitive interactions or PII. Do not follow the commit guidance if the repository has a remote configured you don't trust; ensure no automatic push happens.
- There are no network calls or credential requirements in the included files, so the package does not appear to exfiltrate data by itself — but created files could later be uploaded if you commit or push them. Treat generated logs as potentially sensitive.
- The code uses macOS-style sed -i '' in one script (introspection-prompt.sh) and assumes certain utilities (tree command used in integration script) that may not exist or behave differently on other OSes; test in an isolated environment first.
- If you intend to use this, run the scripts in a dedicated sandbox workspace (empty git repo without remotes) and inspect the files generated before incorporating into any production environment.
If you want, I can (a) produce a safe walkthrough to adapt the scripts to a custom workspace path and remove git suggestions, or (b) analyze any particular script or line you are unsure about in more detail.Like a lobster shell, security has layers — review code before you run it.
latestvk9720mv6md43cswtp42m7hsqyh80j9tx
License
MIT-0
Free to use, modify, and redistribute. No attribution required.