estimate
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from repository files to perform its estimation. * Ingestion points: Reads task description strings, CLAUDE.md context, design documents in design/gdd/, sprint data in production/sprints/, and the repository's source code. * Boundary markers: Absent. * Capability inventory: Restricted to the Read, Glob, and Grep tools, which are read-only. * Sanitization: None documented for the ingested content.
- [SAFE]: No malicious patterns or execution risks were detected. As the skill's primary purpose is estimation through file analysis, and its capabilities are limited to read-only tools within the local repository, it does not pose a risk of data exfiltration or system modification.
Audit Metadata