Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Exec Gilles
v1.0.0Assistant exécutif de haut niveau et partenaire de responsabilisation intraitable pour cadre dirigeant avec TDAH. Gère proactivement Google Agenda, Google Ta...
⭐ 0· 65·1 current·1 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The SKILL.md describes direct, active management of Google Calendar, Google Tasks, Gmail, Google Drive and WhatsApp (including scanning and modifying data). Yet the skill metadata declares no required environment variables, no credentials, and no install/integration spec. A skill that actually performs those actions would need OAuth tokens, API keys, or explicit integration instructions; their absence is incoherent.
Instruction Scope
Runtime instructions tell the agent to proactively scan emails, WhatsApp conversations, and Drive files; to extract commitments and deadlines; and to create/modify/delete calendar events and tasks. Those instructions involve reading and writing sensitive user data and continuous monitoring — all of which expand the scope well beyond a passive helper. The SKILL.md also references a specific WhatsApp phone number, which is unusual and unexplained.
Install Mechanism
There is no install specification and no code files (instruction-only). This reduces the risk of arbitrary code being written/executed on disk, but means integrations must be provided at runtime; the absence of declared integration steps or credential requirements is the core inconsistency.
Credentials
The skill clearly requires access to multiple sensitive services (Google accounts, Drive, Gmail, WhatsApp) but declares no environment variables, primary credential, or config paths. Expected credentials (OAuth tokens, API keys, WhatsApp integration/session data) are missing from the metadata, which is disproportionate and unexplained.
Persistence & Privilege
always is false and the skill is user-invocable (normal). The SKILL.md’s statement that the agent should 'observe its own functioning' and 'propose updates to its own rules' is higher-level behavior guidance but does not by itself indicate installation of persistent privileges. Combine this with the other concerns (broad data access without declared creds) before granting persistent or autonomous access.
What to consider before installing
Do not enable or grant broad access yet. Ask the publisher to explain, in plain terms: (1) how the skill will obtain access to Google Calendar/Tasks/Gmail/Drive and WhatsApp (OAuth flows, redirect URIs, scopes requested), (2) whether it requires any environment variables or tokens and where those are stored, (3) what the referenced WhatsApp number represents and why it is listed in the instructions, and (4) how data access is audited, limited, and revoked. If you test it, use disposable or low-privilege accounts, restrict OAuth scopes to the minimum, review activity logs, and avoid granting write/delete permissions until you’re confident. The current package is missing the integration/credential details necessary to trust it.Like a lobster shell, security has layers — review code before you run it.
latestvk97cf2cthzvv86n0n3y20wgmdh83m1n8
License
MIT-0
Free to use, modify, and redistribute. No attribution required.