The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. It retrieves data from the internet via web_search and web_fetch to generate financial reports. If the retrieved web content contains malicious instructions, the agent may follow them, potentially biasing the analysis or performing unintended actions. Ingestion points: web_search and web_fetch are used in SKILL.md to collect company data and news. Boundary markers: The instructions do not provide boundary markers or directives to ignore instructions embedded in the retrieved data. Capability inventory: The skill is designed to synthesize data into reports. Although SKILL.md explicitly instructs the agent not to write files, the provided skill package includes an output/ directory with generated Markdown reports, indicating file writing capabilities have been exercised. Sanitization: No sanitization or filtering of external content is present.
[EXTERNAL_DOWNLOADS]: The skill uses web_fetch to retrieve content from arbitrary URLs and web_search for broader information gathering. This functionality involves processing unverified external data.
[COMMAND_EXECUTION]: The skill invokes external tools like finx gildata-aidata and asset-service using command-line syntax to access professional financial databases.

company-deep-analysis