The Agent Skills Directory

EXTERNAL_DOWNLOADS (MEDIUM): The installation instructions in README.md suggest using npx skills add KennethAshley/bittensor-cli-skill. This source is an untrusted third-party repository outside the permitted trust scope, posing a supply chain risk.
COMMAND_EXECUTION (HIGH): The skill grants the agent the ability to execute shell commands (python scripts/metagraph.py) and perform sensitive CLI operations (btcli). Combining these capabilities with the ingestion of external, user-controlled data from a decentralized network creates a significant security risk.
Indirect Prompt Injection (HIGH): The skill processes data from the Bittensor metagraph (e.g., validator names and subnet metadata) which is controlled by external network participants.
Ingestion points: Data from the decentralized network enters the agent context through metagraph.py and subnets.py.
Boundary markers: Absent. There are no delimiters or instructions provided to the agent to ignore potentially malicious content embedded in network data.
Capability inventory: The agent has access to script execution and the btcli management tool for wallet and registration actions.
Sanitization: Absent. The skill does not implement validation or escaping for data retrieved from the Bittensor network, allowing potentially malicious strings to be treated as instructions.

bittensor