Skills
Navigate
skills/ebursztein/sprinty/how-to-run-a-sprint/Gen Agent Trust Hub

how-to-run-a-sprint

Pass

Audited by Gen Agent Trust Hub on Oct 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The sprint_close tool accepts a command argument within the coverage configuration object, which allows for arbitrary shell command execution for coverage reporting.
  • [DYNAMIC_EXECUTION]: The skill utilizes 'executable gates' defined within the sprint ledger. These scripts or commands are automatically executed by the item_done tool during task verification.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes metadata and instructions from a persistent disk-based ledger, creating an attack surface for indirect injection. 1. Ingestion points: info, sprint_resume, and search tools reading state from the .sprinty directory. 2. Boundary markers: No explicit delimiters or warnings for instructions embedded in task data are provided. 3. Capability inventory: sprint_close (shell command execution) and item_done (executable gate verification). 4. Sanitization: The instructions do not define sanitization or escaping protocols for data retrieved from the ledger.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 18, 2026, 06:42 PM