The Agent Skills Directory

[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it processes untrusted data from external sources (emails, GitHub issues, and social media posts) to generate its monologue.
Ingestion points: Data enters the context through WorkIQ-ask_work_iq, github-mcp-server-search_pull_requests, and X-twitter-getUsersPosts as described in SKILL.md.
Boundary markers: Absent; the skill does not define specific delimiters or instructions to treat the retrieved content as untrusted.
Capability inventory: The skill's capabilities are limited to text generation. It does not have tools for file system modification, arbitrary command execution, or unauthorized network communication.
Sanitization: No explicit sanitization or filtering of the retrieved data is performed before it is used to generate the roast.
[DATA_EXFILTRATION]: While the skill accesses sensitive user data (emails, meetings, private commits), it does so within the scope of its defined purpose to provide a service directly to the user. No instructions were found that attempt to exfiltrate this data to external or non-whitelisted domains.
[COMMAND_EXECUTION]: The skill utilizes standard tool calls for data retrieval. There are no patterns suggesting shell injection or the execution of arbitrary system commands.

comedy-roast