The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides instructions to install the s1 CLI by piping a remote script directly into the shell (curl -fsSL https://cli.search1api.com/install.sh | bash), which allows for unverified code execution from an external source.
[EXTERNAL_DOWNLOADS]: The skill fetches content and executable code from several external sources, including the cli.search1api.com domain and the npm registry.
[COMMAND_EXECUTION]: The skill's entire functionality relies on executing shell commands via the s1 tool and npx, which involves high-level interaction with the operating system.
[DYNAMIC_EXECUTION]: The s1 learn command dynamically generates and installs new AI agent skills into local directories (~/.agents/skills/ or ./.agents/skills/), which effectively creates new executable content at runtime based on external data.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process arbitrary data from search results and web pages, creating a significant vulnerability to instructions hidden in external content.
Ingestion points: Data enters the context from s1 search, s1 crawl, and s1 learn as defined in SKILL.md.
Boundary markers: No delimiters or instructions to disregard embedded commands are present in the processing logic.
Capability inventory: The agent can execute shell commands and modify the local file system by installing new skills, as shown in SKILL.md.
Sanitization: There is no documented logic for sanitizing or escaping the content retrieved from external URLs before it is processed by the agent.

search1api