start-session

Pass

Audited by Gen Agent Trust Hub on Nov 8, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the repository it is managing, which creates a surface for indirect prompt injection.
  • Ingestion points: The agent is instructed to read contents from AGENTS.md, CLAUDE.md, file-based handoffs, and external tracker items (GitHub issues and pull requests).
  • Boundary markers: No specific delimiters or instructions to ignore embedded commands within the ingested data are defined.
  • Capability inventory: The skill has the ability to execute shell commands via git and gh, and can trigger local framework scripts.
  • Sanitization: There is no evidence of sanitization or filtering for the content retrieved from these external sources before it is used to influence the agent's belief state and task selection.
  • [DYNAMIC_EXECUTION]: The skill triggers the execution of a local shell script framework/skills/_shared/scripts/check-skill-shadowing.sh using a variable repository path argument. This represents the execution of scripts that are not part of the immediate skill body.
  • [DATA_EXPOSURE]: The skill reads from a specific local system path /home/ya/devd/technology_stack to verify the platform's technology stack. This constitutes access to platform-level configuration data to determine the environment's capabilities.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 8, 2026, 11:07 PM