The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill documentation and fallback logic provide instructions to download and execute code from an external repository using uvx --from git+https://github.com/AI-Substrate/flow_squared fs2 install. This pattern involves running unverified third-party software as part of the tool's installation and setup process.
[COMMAND_EXECUTION]: The skill utilizes shell-equivalent tools including Glob, Grep, and Read to perform discovery and content extraction from the local filesystem.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection.
Ingestion points: The agent reads content from files within the target codebase, including docs/domains/registry.md and source code matching query patterns, which may contain attacker-controlled data.
Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands when the agent processes or summarizes findings from these files.
Capability inventory: The skill has capabilities to read files and perform network search operations via the FlowSpace MCP tools.
Sanitization: File content is interpolated into the agent's research findings without visible sanitization or validation logic.
[EXTERNAL_DOWNLOADS]: The skill references and encourages the use of external graphs and semantic search capabilities that may require external API configurations (Azure/OpenAI) and downloads from the AI-Substrate/flow_squared repository.

flowspace-research-v2