ClawHub

Memory Distiller

v0.1.0

Distill repeated user preferences, successful patterns, and durable working rules into reusable memory notes or prompt-ready context blocks. Use when a user...

⭐ 0· 120·0 current·0 all-time
by@danxbuidl
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
✓
Purpose & Capability
The name/description (distilling durable preferences and working rules) matches the instructions and included reference docs. There are no extraneous environment variables, binaries, or config paths requested that would be unrelated to the stated purpose.
✓
Instruction Scope
SKILL.md limits work to user-provided source material (conversations, task outcomes, prior notes) and gives clear rules for evidence thresholds and what to exclude (temporary state, one-offs, guesses). It does not instruct reading system files, environment variables, or contacting external endpoints. It explicitly forbids hidden/background memory injection in runtime code paths.
✓
Install Mechanism
There is no install spec and no code files to write or execute; the skill is instruction-only, which minimizes on-disk risk.
✓
Credentials
The skill declares no required credentials, env vars, or config paths. The instructions do not request secrets or unrelated credentials. The requested scope (user-provided text) is proportionate to the goal of creating memory notes.
✓
Persistence & Privilege
The skill is user-invocable and not marked always:true. Registry metadata allows model invocation by default, but the included agents/openai.yaml sets policy.allow_implicit_invocation: false (reducing implicit/autonomous invocation). There is no instruction to modify other skills or system-wide settings.
Assessment
This skill appears internally consistent and low-risk: it only needs the conversation or notes you explicitly provide and asks for no credentials or installs. Before using it, avoid feeding sensitive secrets or full transcripts containing credentials (because distilled memories could capture them). Review distilled outputs before persisting them in any external memory store, and test with non-sensitive examples first to confirm the format and conservatism meet your expectations.

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

latestvk976dr0mamtvc49n1jj1bmyqns834svh

License

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

Comments