The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructions in SKILL.md and references/deps-dev.md repeatedly direct the agent to "Do not read script source code." This directive restricts the agent's transparency and oversight of its own functional components.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface by ingesting untrusted external data into the agent's context.
Ingestion points: Documentation indices and site content fetched via llms-probe.sh and WebFetch, and source code files cloned via git-clone.sh.
Boundary markers: Absent; the instructions do not implement delimiters or safety warnings for the agent to ignore instructions embedded in the external content.
Capability inventory: The skill can execute shell scripts, Python scripts, perform network operations via curl and git, and run Node.js packages via bunx.
Sanitization: Absent; content from external sources is processed without documented validation or filtering.
[COMMAND_EXECUTION]: The script scripts/llms-probe.sh utilizes the -k (insecure) flag with curl, which bypasses SSL certificate validation. This practice is a security best-practice violation that could expose the agent to potential man-in-the-middle attacks when fetching documentation from remote servers.
[EXTERNAL_DOWNLOADS]: The skill performs remote operations including cloning public git repositories, fetching documentation files via curl, and executing the ctx7 documentation search utility via bunx.

godfetch