The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The file references/integrations.md contains multiple instructions to download and execute remote shell scripts by piping curl output directly into bash or sh. Targets include scripts from astral.sh, opencode.ai, and openclaw.ai.
[COMMAND_EXECUTION]: The documentation provides numerous shell commands for the user or agent to execute, including npm install, docker compose, git clone, and various uv or pip operations across different files such as integrations.md and quick-start.md.
[DYNAMIC_EXECUTION]: In references/integrations.md, there are instructions to dynamically generate a shell script (pcodex.sh) in the user's home directory and set it as executable using chmod +x.
[INDIRECT_PROMPT_INJECTION]: The skill uses WebFetch to dynamically ingest content from external Markdown files hosted on GitHub (raw.githubusercontent.com/jiekouai/jiekou-skills).
Ingestion points: External references for guides and FAQs in SKILL.md.
Boundary markers: No specific delimiters or safety instructions are used to isolate the fetched content.
Capability inventory: The skill has broad capabilities including command execution and network access.
Sanitization: No sanitization is performed on the externally fetched documentation content.

jiekou-docs