The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to fetch documentation from OpenAI's official developer portal and search GitHub for specific protocol RFCs and examples. These targets represent well-known and trusted services.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests untrusted data from the web and maintains powerful tool capabilities.
Ingestion points: Content retrieved via WebFetch from developers.openai.com and WebSearch results from github.com.
Boundary markers: Absent. The instructions do not include delimiters or warnings to ignore malicious instructions that might be embedded in the external documentation.
Capability inventory: The skill has access to Bash for command execution and Write/Edit for file system modifications.
Sanitization: Absent. There are no mechanisms described to validate or filter the content retrieved from external sources before it is processed by the agent.

acp-extensions-authoring