daishin-report-search
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill relies on executing commands via
npxto fetch instructions, list files, and perform updates using the vendor's package (@nomadamas/k-skill). - [EXTERNAL_DOWNLOADS]: The skill performs automated network requests to GitHub API endpoints and raw content servers (
api.github.com,raw.githubusercontent.com) to retrieve financial report data from a third-party repository. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted HTML content from an external source, creating a potential vector for instructions embedded in reports to influence the agent.
- Ingestion points: HTML report files and accompanying explanation pages retrieved from the
jay-jo-0/github_pages_reporepository. - Boundary markers: The skill instructions suggest conservative summarization but do not implement strict delimiters or instructions to ignore potential commands within the fetched data.
- Capability inventory: The skill possesses network read capabilities to fetch data and outputs processed text for agent consumption.
- Sanitization: There is no evidence of sanitization or filtering to remove potential prompt injection attempts from the ingested HTML content.
Audit Metadata