The Agent Skills Directory

[SAFE]: The skill is entirely composed of Markdown-based instructions and checklists for analytical reasoning. It does not contain any scripts, binary files, or configuration that would allow for command execution or system modification.
[SAFE]: No network operations (such as curl, wget, or fetch) or hardcoded credentials were detected. The skill specifically instructs the agent not to fetch live data unless explicitly requested by the user through a different skill.
[SAFE]: There are no signs of obfuscation, hidden URLs, or malicious redirection patterns. The content is written in clear, domain-specific language related to financial analysis.
[INDIRECT_PROMPT_INJECTION]: While the skill is designed to process untrusted user input (investment theses and supporting notes), it lacks the dangerous capabilities (like file system writing, network requests, or shell execution) that would make such an injection exploitable. The instructions focus on logical reasoning and categorization of the provided text.

thesis-validation