The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes ripgrep (rg) to discover project structures by searching for configuration files and curl to verify that generated favicons are correctly deployed and accessible via network requests. These commands are consistent with the skill's stated purpose of project automation and validation.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the project environment that could potentially influence the agent's behavior.
Ingestion points: The skill reads package.json, README.md, and site homepages to determine site semantics and design direction, as specified in the '发现站点与配置' and '选择图标语义' sections of SKILL.md.
Boundary markers: Absent; the instructions do not provide delimiters or specific warnings to the agent to disregard potential instructions embedded within the ingested project files.
Capability inventory: Filesystem searching (rg), file creation and modification (writing .svg files and updating .ts/.js configuration files), and network reading (curl).
Sanitization: Absent; there is no mention of sanitizing or validating the strings extracted from project documentation before they are used to influence the agent's design and configuration tasks.

add-favicon