The Agent Skills Directory

[SAFE]: The skill implements a standard cryptocurrency monitoring utility using public APIs from reputable providers including Whale Alert, CoinGecko, and Etherscan.
[SAFE]: Data persistence is handled via local JSON files in the user's home directory (e.g., ~/.whale_watchlist.json), which is standard behavior for CLI-based monitoring tools to maintain state between sessions.
[SAFE]: Tool permissions are restricted via YAML frontmatter using a pattern-matched Bash tool (Bash(python:*whale*)), limiting the execution scope to the intended scripts.
[SAFE]: Analysis of the Python scripts confirms they perform legitimate data retrieval, formatting, and local storage operations with no signs of obfuscation, exfiltration, or unauthorized command execution.
[PROMPT_INJECTION]: Indirect Prompt Injection Surface:
Ingestion points: Data is ingested from the Whale Alert API in scripts/whale_api.py and CoinGecko API in scripts/price_service.py.
Boundary markers: None explicitly defined in the scripts, though output is primarily formatted into tables and JSON.
Capability inventory: The agent has access to file system tools (Read, Write, Edit, Grep, Glob) and restricted Bash execution.
Sanitization: The scripts perform standard data type conversion and truncation (in scripts/formatters.py) but do not explicitly filter for prompt injection markers in API-returned strings (e.g., wallet labels or owner names). This is a common surface for data-processing skills and represents a low risk in this context.

monitoring-whale-activity