The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute multiple command-line utilities for documentation lookup, project environment inspection, and skill maintenance. These include the tanstack CLI (search-docs, doc, libraries), the opensrc utility for locating source files, the bun package manager, and local Python scripts located at tools/skill/quick_validate.py and tools/skill/check_tanstack_skills.py.
[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by design, as it instructs the agent to ingest and act upon data retrieved from external sources at runtime.
Ingestion points: Data enters the agent's context through the output of tanstack search-docs ... --json and by reading source code files directly from the node_modules directory.
Boundary markers: The instructions lack defined delimiters or warnings to the agent to disregard potential instructions embedded within the external documentation or source code it retrieves.
Capability inventory: The skill environment grants the agent the ability to execute shell commands (python3, bun, tanstack) and perform file system reads, which could be exploited if malicious content is ingested.
Sanitization: There are no requirements or procedures defined for validating or sanitizing the content retrieved from external tools before the agent processes it for decision-making.

tanstack-integration