fittings-quake-nearby
Warn
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: MEDIUMCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The skill downloads a JavaScript file from a remote vendor URL (https://fittings.sh/skill/call.mjs) and executes it using the Node.js runtime. This bypasses local skill content and introduces a dependency on external code fetched at execution time.
- [DATA_EXPOSURE_AND_EXFILTRATION]: The documentation instructs the agent to store a wallet private key in the
FITTINGS_PRIVATE_KEYenvironment variable. The skill explicitly notes that the key sits in plaintext where the agent runs, which is a high-risk credential handling practice. - [COMMAND_EXECUTION]: The skill facilitates the execution of shell commands, including
curlfor fetching remote assets andnodefor running the metered payment client. - [INDIRECT_PROMPT_INJECTION]: The skill processes external parameters that are passed directly to a shell-executed script. 1. Ingestion points: lat, lon, radiusKm, days parameters in SKILL.md. 2. Boundary markers: None present. 3. Capability inventory: execution of node call.mjs. 4. Sanitization: No explicit sanitization or validation of input types is mentioned before passing to the shell.
Audit Metadata