The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external SKILL.md files within the repository to populate the project's README. This creates a surface where malicious instructions embedded in a skill's description could be propagated into the main documentation.
Ingestion points: The process involves scanning all skill/module directories and reading their SKILL.md or equivalent descriptor files (referenced in SKILL.md step 1).
Boundary markers: There are no instructions to use specific delimiters or warnings to ignore potentially malicious embedded instructions in the ingested data.
Capability inventory: The agent utilizes Edit and Write tools to modify the root-level README.md and README.zh-CN.md files.
Sanitization: The instructions do not specify any sanitization, escaping, or filtering of the content read from the repository before it is written to the README files.

update-readme