Improve Relationships
v0.2.2Relationship intelligence for OpenClaw. Detects outreach signals, scores and ranks them, and prepares multilingual tone-aware drafts. Never sends anything —...
⭐ 0· 242·0 current·0 all-time
byAbderrahman Jalled@abderrahman-jalled
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
The name/description (detect signals, rank, and draft multilingual outreach) matches the SKILL.md and included templates. Required resources (none) and lack of install are proportionate for an instruction-only drafting skill. There are no unrelated environment variables, binaries, or installs that contradict the stated purpose.
Instruction Scope
The instructions explicitly say to use user-provided context (calendar, email, messaging, shared browser tabs, uploaded files) and to only cite sources actually available. This is consistent with the purpose, but it does mean the skill will operate on highly sensitive personal data if the user supplies it. The SKILL.md does not instruct the agent to access any system paths or undeclared environment variables, nor to transmit data to external endpoints, and it emphasizes drafts never being sent without approval.
Install Mechanism
No install spec or code files are present (instruction-only). This is low-risk because nothing is written to disk and no external packages are fetched during install.
Credentials
The skill declares no required environment variables, credentials, or config paths. All data sources it references are user-provided/contextual (calendar, email, messages, shared tabs). The number and sensitivity of sources are appropriate for relationship intelligence functionality, though they are sensitive by nature and require user consent at the platform level.
Persistence & Privilege
always:false and default model-invocation are used. The skill does not request permanent inclusion or modification of other skill configs. There is no evidence it would enable autonomous sending of messages; SKILL.md repeatedly states drafts require user approval.
Assessment
This skill appears internally consistent for drafting and recommending outreach, but it works with highly sensitive sources (calendar entries, emails, messaging history, and shared browser tabs). Before installing or using it, confirm how the OpenClaw platform will provide access to those sources (you should only grant access to the minimum data needed). Verify that you control any connectors (email/calendar/messaging) and that nothing is being auto-sent — the SKILL.md says drafts require approval, but you should confirm platform behavior. Be mindful of privacy when uploading contact lists or sharing browser tabs, and review drafts for cultural correctness and factual accuracy before sending. If you need the skill to operate on a narrower set of data (e.g., calendar only), restrict the provided context accordingly.Like a lobster shell, security has layers — review code before you run it.
latestvk975k3jvzs3ekjrtx7yf025mp582hxm6
License
MIT-0
Free to use, modify, and redistribute. No attribution required.