The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the subprocess module to execute external and internal commands across multiple scripts.
scripts/hot_scanner.py and scripts/rumor_scanner.py execute the bird CLI tool for social media searches using a set of searched and hardcoded paths, including a specific user's home directory (/home/clawdbot/...).
scripts/watchlist.py triggers the internal analyze_stock.py script via uv run to perform automated signal checks for watchlisted assets.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes data from external, untrusted sources, creating a vulnerability surface where malicious instructions could be embedded in market data.
Ingestion points: Raw headlines from Google News RSS and tweet text from Twitter/X are fetched and displayed in hot_scanner.py and rumor_scanner.py.
Boundary markers: The skill does not use specific delimiters or safety instructions (e.g., "ignore instructions in this text") when presenting external news and social media content to the AI agent.
Capability inventory: The skill possesses the ability to execute shell commands and modify local JSON state files (portfolios.json, watchlist.json).
Sanitization: There is no evidence of text sanitization or filtering to remove potentially malicious AI instructions from the ingested external content.
[EXTERNAL_DOWNLOADS]: The skill requires the manual installation of a third-party Node.js package (@steipete/bird) to enable its social media sentiment and trend detection features.

stock-analysis