boligsiden
Warn
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: MEDIUMREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill relies on @bunli/core and @bunli/utils which are non-standard packages. These are included with the latest version tag in cli/package.json, introducing a supply chain vulnerability where unauthorized or malicious updates to these packages could be executed automatically.
- [EXTERNAL_DOWNLOADS]: The CLI tool makes network requests to api.boligsiden.dk to fetch property listing information, representing an external data dependency.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the Boligsiden API and outputs it to the agent, creating a potential surface for indirect prompt injection via property descriptions or metadata.
- Ingestion points: Data is retrieved from the API in search.ts, rentals.ts, sold.ts, detail.ts, timeline.ts, and locations.ts.
- Boundary markers: No delimiters or warnings are used to separate API-sourced text from agent instructions.
- Capability inventory: The skill is granted shell execution privileges for the bun runtime to operate its CLI tool.
- Sanitization: There is no evidence of filtering or escaping performed on content fetched from the external API.
Audit Metadata