The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources, including web search results and user-generated content (UGC) from platforms like Reddit, Telegram, and Discord.
Ingestion points: SKILL.md (Step 2) and references/cli-reference.md specify the use of news feed web-search, news feed search-news, news feed get-social-sentiment, and news feed search-ugc to collect external data.
Boundary markers: SKILL.md (Step 3: Synthesis) contains instructions for the agent to explicitly separate and label "on-chain / CLI facts" from "community or media interpretation."
Capability inventory: The skill executes a wide range of gate-cli subcommands (e.g., info onchain, info platformmetrics) as detailed in references/cli-reference.md.
Sanitization: The instructions focus on structural separation of data types in the final report rather than automated filtering or sanitization of the input content.
[EXTERNAL_DOWNLOADS]: The skill includes maintenance scripts (scripts/update-skill.sh and scripts/update-skill.ps1) designed to keep the skill up to date.
The scripts fetch the latest version of the skill's code from the vendor's official GitHub repository at github.com/gate/gate-skills using standard tools like git, curl, or wget.
[COMMAND_EXECUTION]: The skill relies on the execution of the gate-cli tool to perform its primary analysis functions, which involves invoking various subcommands to fetch protocol and wallet data.

gate-info-web3