pp-zameen
Pass
Audited by Gen Agent Trust Hub on Jul 27, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to install the
zameen-pp-clitool usingnpx -y @mvanhorn/printing-press-libraryandgo install github.com/mvanhorn/printing-press-library/.... These resources are hosted on the vendor's own infrastructure. - [COMMAND_EXECUTION]: The skill relies on the
Bashtool to executezameen-pp-clifor property searches, data mirroring, and market analysis. - [DATA_EXFILTRATION]: The tool includes a
--deliver webhook:<url>feature that allows routing command output to any external URL provided by the user, which is a functional capability for data transport. - [PROMPT_INJECTION]: The skill ingests untrusted property listing data from Zameen.com search results. This represents an indirect prompt injection surface as the agent processes server-rendered content.
- Ingestion points: Zameen.com search pages processed via
listingsandfindcommands (SKILL.md). - Boundary markers: The skill uses structured JSON output with a provenance envelope and allows the agent to
--selectspecific fields, which helps limit the exposure of untrusted content. - Capability inventory: The skill has
ReadandBashcapabilities and can execute arbitrary CLI commands via thezameen-pp-clibinary. - Sanitization: The skill advises the agent to strip PII (names, emails, phones) before 'teaching' new patterns to the local learning loop.
Audit Metadata