The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill prefers using npx @solana-compass/cli@latest, which causes the environment to download and execute the latest version of the CLI tool from the npm registry whenever a command is run.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources that could be influenced by third parties.
Ingestion points: The sol fetch command (described in SKILL.md and references/fetch-commands.md) retrieves content from arbitrary URLs. Commands like sol token browse and sol predict list ingest data from external trading and prediction market APIs (Jupiter, Polymarket, Kalshi).
Boundary markers: There are no specified boundary markers or instructions to ignore embedded commands within the fetched data.
Capability inventory: The tool possesses high-impact capabilities including token transfers (token send), token swaps (token swap), and staking operations (stake new).
Sanitization: No sanitization or validation of the fetched external content is described before it is processed by the agent.
[COMMAND_EXECUTION]: The skill operates primarily through shell command execution using the sol or npx CLI to perform blockchain transactions.
[SAFE]: The skill incorporates a defensive security model with features like fine-grained permissions (e.g., canTransfer, canSwap), transaction USD limits, and allowlists for addresses and tokens to mitigate risks during autonomous operation.
[SAFE]: The documentation provides proactive security advice, specifically warning the user not to grant the agent direct file system access to the ~/.sol/ directory where private keys are stored.

solana-payments-wallets-trading