planning-suite-reconciler
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches skill components from the author's official GitHub repository (github.com/peter-tu-zynkr/zynkr-skill-builder) during installation via npx.\n- [SAFE]: Adheres to a strict 'read-only' policy for the source-of-record spreadsheet and mandates a single user confirmation before executing any batch write operations to Google Docs or Sheets.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the Main Tracker to generate document updates. \n
- Ingestion points: Data is read via read_sheet_values from the tracker's SOR and retro tabs (Step 1).\n
- Boundary markers: Uses a fixed heading structure and a mandatory conflict disclaimer sentence defined in the knowledge pack.\n
- Capability inventory: Performs file-write operations using batch_update_doc and modify_sheet_values (Steps 6 and 7).\n
- Sanitization: Instructions require stripping markdown markers from ingested text before insertion into documents to prevent formatting-based injection.
Audit Metadata