The Agent Skills Directory

[SAFE]: The skill functions as a blockchain monitoring and analysis tool, utilizing well-known industry APIs for data retrieval without any detected malicious intent, obfuscation, or harmful operations.
[EXTERNAL_DOWNLOADS]: The skill fetches blockchain transaction signatures, parsed transaction data, and market analytics from reputable providers including Helius (api.helius.xyz), SolanaTracker (data.solanatracker.io), and Birdeye (public-api.birdeye.so). These operations are essential for its purpose and directed at legitimate services.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data from the Solana blockchain. 1. Ingestion points: Fetches real-time transaction history and holder snapshots in scripts/track_whales.py and scripts/whale_alerts.py. 2. Boundary markers: Data is processed as informational values within scripts; there are no explicit instructions for an agent to interpret embedded data as commands. 3. Capability inventory: Limited to network operations via the httpx library and terminal output; the skill lacks capabilities for file system modification, subprocess spawning, or persistent access. 4. Sanitization: Scripts use structured data models to parse and extract specific numerical and address fields from raw JSON responses, which inherently filters out potential injection payloads in unexpected fields.

whale-tracking