The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes several project-specific commands and build tools. These include cargo xtask research scaffold and cargo xtask research check for folder management and validation. It also executes a suite of documentation tools via bun (build, check:repo-links, tsc, and tests) and standard git commands for committing and pushing research dossiers.
[DATA_EXFILTRATION]: The skill performs legitimate data transfers consistent with its purpose. It gathers information from external sources using a built-in deep-research tool and pushes generated documentation to the vendor's repository. These operations are essential to the primary research and publishing function.
[PROMPT_INJECTION]: The skill is subject to indirect prompt injection risks as it ingests and synthesizes untrusted data from the web and the codebase.
Ingestion points: External data is ingested via the deep-research tool; local codebase data is ingested via Explore and grep operations.
Boundary markers: None explicitly defined in the instruction file to isolate the synthesized content from the agent's instructions.
Capability inventory: The agent has the capability to write to the local filesystem (docs/content/...), execute shell commands (bun, cargo, git), and push changes to a remote repository.
Sanitization: The skill does not mention specific sanitization or escaping of external content before processing, relying on the agent's internal safety guardrails.

research