ClawHub

🤖🤝🧠 better collab with your agent

v1.0.0

Analyze ChatGPT conversation exports to discover cognitive archetypes and optimize AI-human communication patterns. Enables personalized agent interactions based on detected user profiles.

⭐ 3· 1.7k·2 current·2 all-time
bySebastian Otálora@sebastianffx
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
medium confidence
ℹ
Purpose & Capability
The name/description match the code: scripts analyze ChatGPT conversation exports and produce a profile to tune agents. Requesting python3 is appropriate. Minor inconsistency: SKILL.md/README state that only the standard library is required (scikit-learn/numpy are listed as recommended), but requirements.txt lists scikit-learn, numpy, and PyYAML as core dependencies — unclear which packages are actually required for normal runs.
ℹ
Instruction Scope
Primary runtime instructions (export ChatGPT data, run scripts/analyze_profile.py) are coherent and limited to local processing of conversations.json. However: (1) the package includes test_wildchat.py which instructs downloading/streaming the external 'allenai/WildChat-1M' dataset (network activity and third‑party data access) but SKILL.md doesn't mention this test script; (2) the scripts write output to a home config path (~/.openclaw/...), which the metadata did not declare as a required config path. These are scope surprises rather than obvious malicious behavior.
ℹ
Install Mechanism
No install spec in the skill manifest (instruction-only). There is a requirements.txt and README instructs 'pip3 install -r requirements.txt' — that will install scikit-learn/numpy/PyYAML per the file. This is a normal Python dependency install, but the manifest's claim that only the standard library is required contradicts requirements.txt. No remote or obfuscated installers were found; the package appears to rely on standard PyPI packages.
✓
Credentials
The skill does not request environment variables or credentials. It processes local ChatGPT exports and generates a JSON profile. One proportionality note: scripts and SKILL.md write output to ~/.openclaw/my-cognitive-profile.json and allow creating ~/.openclaw/my-archetypes.yaml — the manifest did not list required config paths, so users should be aware the tool will create/read files in the user's home directory.
✓
Persistence & Privilege
The skill does not request always:true, elevated privileges, or modify other skills. It is user-invocable and can be run locally. Autonomous invocation is allowed by default for skills on the platform, but nothing in the package indicates forced or hidden persistence.
What to consider before installing
This package mostly does what it says — it analyzes your ChatGPT export locally and produces a profile — but there are a few things to check before running: 1) Review requirements.txt vs SKILL.md: decide whether you want to install scikit-learn/numpy/PyYAML (they are listed as core deps in requirements.txt but described as optional/recommended in the docs). 2) Be cautious with test_wildchat.py — it loads a large external dataset (allenai/WildChat-1M) via the datasets library and will perform network I/O; don't run it if you want strictly local processing. 3) The tool writes output and may create ~/.openclaw/* files; inspect those paths or change them to a location you control. 4) Inspect the scripts if you have sensitive data: the analyzer code appears local (no hidden remote endpoints), but always examine third-party dependencies you install. Recommended: run the analyzer in an isolated virtualenv (or container), review scripts/analyze_profile.py and test_wildchat.py yourself, and avoid running test scripts that pull external datasets unless you intend to.

Like a lobster shell, security has layers — review code before you run it.

archetypesvk976qyzy5yhbabbxtk1b4zhfqs80d1b0chatgptvk976qyzy5yhbabbxtk1b4zhfqs80d1b0clusteringvk976qyzy5yhbabbxtk1b4zhfqs80d1b0communicationvk976qyzy5yhbabbxtk1b4zhfqs80d1b0latestvk976qyzy5yhbabbxtk1b4zhfqs80d1b0nlpvk976qyzy5yhbabbxtk1b4zhfqs80d1b0personalizationvk976qyzy5yhbabbxtk1b4zhfqs80d1b0

License

MIT-0
Free to use, modify, and redistribute. No attribution required.
Termshttps://spdx.org/licenses/MIT-0.html

Runtime requirements

🤖🤝🧠 Clawdis
Binspython3

Comments