Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Garmer
v1.0.2Extract health and fitness data from Garmin Connect including activities, sleep, heart rate, stress, steps, and body composition. Use when the user asks about their Garmin data, fitness metrics, sleep analysis, or health insights.
⭐ 0· 2k·2 current·2 all-time
by@garrza
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, code, and README consistently describe a Garmin Connect data extractor that needs a Garmin account and tokens; that capability aligns with the requested behavior. However, the registry metadata (shown at the top of the submission) lists no required env vars or primary credential while the SKILL.md declares a primaryEnv (GARMER_TOKEN_DIR) and the code/README reference token files (~/.garmer/garmin_tokens) and interactive login. This mismatch (registry claims no credentials required but the skill clearly needs Garmin credentials/tokens) is an inconsistency to be aware of.
Instruction Scope
SKILL.md and examples instruct the agent to run the 'garmer' CLI and Python API to authenticate (interactive email/password via 'garmer login'), read saved tokens from ~/.garmer/garmin_tokens (or a directory set by GARMER_TOKEN_DIR), and fetch health data. The instructions are scoped to the stated purpose (fetching Garmin data) and do not instruct the agent to read unrelated system files or send data to unexpected endpoints. They do, however, require collecting user credentials (email/password) for Garmin and storing tokens locally.
Install Mechanism
No separate install spec was present in the registry metadata summary, but SKILL.md includes install metadata for pip/uv and the package contains a standard Python project (pyproject.toml) and a console script entry point. Installing via pip (from PyPI or local source) is a standard pattern and not inherently high risk; there is no evidence of arbitrary URL downloads or extraction from untrusted servers. The inconsistency between 'no install spec' in the registry summary and install metadata inside SKILL.md is a packaging/metadata mismatch that should be confirmed.
Credentials
The skill needs only Garmin credentials/tokens and optionally a token directory env var (GARMER_TOKEN_DIR) and logging/cache env vars referenced in README. It does not request unrelated cloud credentials or other secrets. That is proportionate to its purpose. Again, the registry-level declaration of 'no required env vars' contradicts the SKILL.md/readme which list GARMER_TOKEN_DIR and describe token storage; confirm which declaration is authoritative.
Persistence & Privilege
The skill does not request always: true and does not attempt to modify other skills or global agent config. It stores OAuth tokens under ~/.garmer/garmin_tokens (or a custom directory) for reuse; this is expected for a client that needs persistent auth. Note that autonomous invocation is enabled by default (disable-model-invocation: false) — this is platform normal behavior, but you should be comfortable with an installed skill being invoked by the agent when appropriate.
What to consider before installing
This package appears to be a real Garmin Connect data extractor, but the registry metadata and the SKILL.md/README disagree about what is required. Before installing or using it: 1) Confirm the package source and author (the registry shows an anonymous owner id and no homepage). 2) Inspect the code (you have the source here) or run it in an isolated environment if you will provide your Garmin credentials. 3) Understand that you will need to enter your Garmin email/password during 'garmer login' and that tokens will be stored under ~/.garmer/garmin_tokens or a directory set via GARMER_TOKEN_DIR — treat those tokens as sensitive and revoke them in your Garmin account if needed. 4) Prefer installing from a trusted package index or from source you control, and consider running the tool inside a VM or container if you are unsure about provenance. If you want, I can highlight any code files or functions to review first (authentication flow, network endpoints used, and token storage/refresh logic).Like a lobster shell, security has layers — review code before you run it.
garminvk978xvppm2r8m8cdphf9pwgrz980gjrslatestvk973rmbjzg5pqaf424ydxwdr3s80gc1vlatest, garmin,health,wellness,garmervk978xvppm2r8m8cdphf9pwgrz980gjrslatest,garmin,health,wellness,garmervk978xvppm2r8m8cdphf9pwgrz980gjrs
License
MIT-0
Free to use, modify, and redistribute. No attribution required.